CVIS Research Blog

Recruitment for RA for Computer Vision and Image Processing project

2012-01-12T17:39:00.002+08:00

Dear all,

Prof Maylor Leong has a vacancy for research assistant (RA) for 12 months to work in the area of computer vision and image processing.
Please contact Prof Maylor Leong at leungkh@utar.edu.my for further details.

Topic: Object recognition employing shape based indexing
Abstract:
The aims of this proposal are two-folds: to research on an object finder for the visually impaired (VI) and to guide the VI to reach such an object, making use of the Computer Vision technique.
For a few decades, computer scientists and engineers have attached cameras to a computer and attempt to teach a machine to see. Great success has been achieved in controlled environment when the lighting and background can be controlled. The problem remains unsolved in uncontrolled places, in particular, when objects are placed in arbitrary poses in cluttered and occluded environment. In general, object recognition is a difficult process since it needs to recognize object category. For example, a car can come with different makes and shapes. To make the matter worse, a car can also come with particular painting or pattern on its body. The recognition of a particular car is not very interesting while the recognition of the car category is more meaningful and can find more useful applications. Researchers have employed learning technique to train a classifier to tackle this category recognition problem in controlled environment. However, few have tackled this problem in cluttered environment. So far, no effective solution has yet been found.
According to our aims, we propose to research on an object finder trained to recognize at least 20 common types of objects found in cluttered environment. The proposed guidance for the VI to reach an object should investigate the matching process from object physical space to the mental space of the VI. An intelligent HCI (human computer interface) employing multi-media processing is needed. For the recognition part, there are many good approaches for single object recognition in controlled environment but there is no working technique for multiple objects recognition in cluttered environment. Here, the proposed approach is based on the idea that human can identify and recognize an occluded object from a cluttered environment, provided that enough locally perceived clues and/or distinguishable portion of the object can be observed. This observation forms the core idea for this work to test and research on. The proposed work is relevant to world emphasis on assistive technology for the elderly and less fortunate. Furthermore the researched techniques can be readily employed on a robot to support work on service robot for elder care and domestic task. This is relevant and significant to industry in Malaysia.

CCIS Seminar: Image Super Resolution: Its Techniques and Applications

2011-12-12T14:23:00.001+08:00

Dear all,

You are welcome to attend the following talk:

Date: 14 December 2011 (Wednesday)
Time: 11am - 12pm
Venue: MIMOS Lab, Ground Floor, SB Block, UTAR KL Campus, Kuala Lumpur.
Speaker: Chua Kah Keong
Title: Image Super-Resolution: Its Techinques & Applications

Abstract:
Super-resolution (SR) image reconstruction is the process of combining several low resolution (LR) images to form a High Resolution (HR) image. In this presentation, the principle of image SR and several techniques are introduced. In addition, the image SR applications and enhancement using Graphical Processing Unit (GPU) will be covered as well.

DMAS & CMS Seminar Series, 9/12/2011, Friday 11am-12pm, Level 6, West Room, SA block, Setapak Campus

2011-12-08T16:42:00.001+08:00

==============================================================================
Department of Mathematical and Actuarial Sciences & Centre of Mathematical
Sciences Seminar Series

Date: 9 December 2011 (Friday)
Time: 11am - 12 pm
Venue: Level 6, West Room, SA block, UTAR Setapak Campus, Kuala Lumpur.
Speaker: Prof. Jeremy Levesley
Title: Radial Basis Function Approximation
===============================================================================

Dear all,

This week's DMAS & CMS talk will be on Friday, 11am-12pm. The speaker is Prof.
Jeremy Levesley from Department of Mathematics, University of Leicester, UK.

Abstract:
Radial basis functions have become an important tool for scattered data approximation, and theoretically can be used for approximation in
high dimensions. There is a very beautiful abstract theory of approximation involving so called native space (which I will not talk about). There is some interest in the finance industry in RBFs because of the possibility of high dimensional approximation and solution of partial differential equations. In this talk I will discuss some of the problems which arise with approximation with such functions, such as ill-conditioning and computation of full linear systems. I will offer some hope for the solution of these problems.

All are welcome.

Thank you very much.

Regards,
Event Committee.

Research Assistant Wanted

2011-12-01T16:28:00.001+08:00

Research Assistant Wanted

The Department of Civil Engineering, Universiti Tunku Abdul Rahman (UTAR), is currently looking for suitable candidates to assist in the Transportation research projects. The research projects are "Microscopic Modeling of Drivers' Merging Behavior at Roundabouts under Oversaturated Condition" by UTAR Research Fund (UTARRF) and "The Novel Development of A Dynamic In-vehicle Guidance and Navigation System in Malaysia" by Exploratory Research Grant Scheme (ERGS). The duration of the appointment is one year. We are looking for highly motivated and dedicated candidates to join us in the research work. The candidates should also keen to register with us as a postgraduate student. The monthly stipend paid to the candidates is RM 2000 (UTARRF) or RM1500 (ERGS). The expected work commencement date is 1^st Jan 2012.

Job Description:

· Traffic data collection

· Process and analyse the collected data.

· Develop mathematical models to describe drivers' behavior or willing to pay for travel information.

· Develop an intelligent route searching algorithm or traffic simulation model for driver's behavior.

Requirements:

1. Candidate with Bachelor Degree in Civil Engineering, Mechanical Engineering, Mechtronics Engineering, Electrical Engineering, Computer Science, Physics, Applied Mathematics with Computing are welcome to apply.

2. Candidates should have at least a 3.0 CGPA (4-point system) for their Bachelor Degree study.

3. Candidates with computer programming knowledge are an added advantage.

4. Candidates should register as postgraduate students (Master by Research) with UTAR.

Interested candidates, please contact me via email for the interview arrangement:

Dr. Khoo Hooi Ling

Assistant Professor/Head

Department of Civil Engineering

Faculty of Engineering and Science

Email: khoohl@utar.edu.my

No. Telephone: 03-41079802, Ext. 151

Work Completion Seminar: New SIFT-based Calibration Methods for Hybrid Camera System

2011-11-24T15:25:00.001+08:00

Dear all,

You are most welcome to attend the Work Completion Seminar by Mr Mr. Low Yi Qian (Master of Engineering Science), on his research title of New SIFT-based Calibration Methods for Hybrid Camera System. Details are as follow:

Date: 25/11/2011 (Fri)
Time: 11:00am-12:00noon
Venue: 5A meeting room, 5th floor, SA Block, FES, UTAR KL Campus.
Research title: New SIFT-based Calibration Methods for Hybrid Camera System
Candidate: Low Yi Qian
Abstract:
Multi-cameras networks have an important role in surveillance systems. In traditional cameras networks mainly based on static cameras. Static camera provides low resolution and poor image quality in modeling. At the meanwhile, Pan-Tilt-Zoom (PTZ) camera is able to obtain multi-angles of views and multi-resolution information; it provides more functionality compared with static camera. In multi-camera networks, they obtain two or more different perspectives views of scene to perform calibration within one another. Traditional approach to solve this problem is by using human inputs or referenced objects to find matched correspondences among all images. The recent works by some researchers have brought remarkable increase of automation to these problems. One of the main challenges in multi-cameras calibration is to obtain an accurate and fast estimation of disparities of two different views of scene. Besides disparities, differences might be caused by occlusion of the object, specular reflection, sensor noise and various other causes.
In this project, we designed and constructed a new hybrid camera system instead of using two static cameras. More precisely, our hybrid camera system consists of a static wide angle camera and a PTZ camera. Both cameras obtain different optical elements and resolutions. We proposed a master-slave concept to represent both cameras. A static camera is master camera with wide angle view is used to monitor the environment from a distance. The images taken from the master camera will be used as reference images. When the master camera detects an object of interest, the slave camera - PTZ camera, will zoom into the region of interest (ROI). PTZ camera plays an important role at pointing towards the ROI in high resolution as well as providing different parameters of PTZ values. Then, we studied and proposed calibration methods for the hybrid camera system. In particular, Scale Invariant Features Transform (SIFT) was used to identify keypoints to perform multiple views matching. One of the major contributions of this project is to design and develop a new smart filtering technique based on the principle of trigonometry, called Trigonometry Filter Technique (TFT). TFT is used to filter the false keypoints by calculating and classifying the orientation of the images. Furthermore, TFT also increases the determination and detection rate of ROI image localization. Secondly, we proposed a novel calibration method based on multiple-image stitching approach. By stitching images (by PTZ camera), a high resolution panorama image will be created. Finally, we are able to stitch and calibrate both static master camera and PTZ camera images without the use of any object as reference. By eliminating the use of reference objects, the obtained empirical results show good performance to handle images from various kinds of angles, scaled and target uncertainties.

Marco Tempest: The augmented reality of techno-magic

2011-11-05T13:04:00.002+08:00

IEEE SMC DLP: Humanized Computational Intelligence with Interactive Evolutionary Computation

2011-11-04T01:23:00.004+08:00

You are invite to attend the IEEE SMC Malaysia Section Chapter Distinguished Lecture Seminar: Humanized Computational Intelligence with Interactive Evolutionary Computation conducted by Prof. Hideyuki TAKAGI from Kyushu University, Japan.

Date: Friday, December 09, 2011
Time: 11am – 12pm
Venue: NMES Lab, FOE, MMU Cyberjaya
Topic: Humanized Computational Intelligence with Interactive Evolutionary Computation
Speaker: Prof. Hideyuki Takagi

Your presence for this seminar will be highly appreciated.

Machine Learning Workshop

2011-11-03T16:21:00.001+08:00

UTAR Robotic Society will be organizing a 1-day Machine Learning Workshop on 12 November 2011.

Below are the details of the workshop:

Title: Machine Learning Workshop
Speaker: Tay Yong Haur
Date: 12/11/2011 (Sat)
Time: 9am-5pm.
Venue: OSK Lab, FES, UTAR KL, Setapak, Kuala Lumpur.
Fee: RM15/person (inclusive of refreshment and lunch).

Topics covered are:
-Introduction to ML
-Introduction to Neural Networks
-Hands-on Neural Networks

Please check out the details of registration here: http://utarrobotic.blogspot.com/2011/11/machine-learning-workshop.html

Research Scholar Wanted (Kinect Project)

2011-10-31T12:47:00.001+08:00

Research Scholar Wanted (Kinect Project)

Interested in learning and research? Interested in Kinect development? Interested in getting a postgraduate degree while doing something fun? If you answer 'Yes' to any of these questions, then you are welcome to apply for the position of Research Scholar in the research lab under Faculty of Information Technology at Multimedia University Cyberjaya campus.

Details of the project as follows:

Project Sponsor: Intel Malaysia

Starting Date: immediate

Stipend: RM2,100 / month

Duration: 2 years (subject to review of performance at end of 1st year)

Field of research: image processing / computer vision / Kinect development / computer graphics / natural gesture interaction

Requirement: degree related to Computer Science / IT / Applied Maths, strong in programming (C++ preferred), strong in mathematics

Location: Faculty of Information Technology, Multimedia University, Cyberjaya

Project Leader: Wong Ya Ping

Contact:

If you are interested, please send your CV to:
"Wong Ya Ping" <wongyaping@gmail.com>
with the subject line : [RS Application] <your name>

Thank you.

Hong Kong PhD Fellowship Scheme 2012/13

2011-09-29T23:10:00.002+08:00

If you are interested to pursue your PhD in universities in Hong Kong, don't miss the opportunity for the PhD Fellowship Scheme available here: http://cerg1.ugc.edu.hk/hkpfs/index.html

Due dates for the application submission is on early December 2011.

CCIS Seminar: Robust Fitting and Model Selection in Computer Vision

2011-09-27T00:56:00.002+08:00

Dear Colleagues,

Centre for Computing and Intelligent Systems (CCIS), Universiti Tunku Abdul Rahman (UTAR) will be organizing the following open seminar:

Date: 7 Oct 2011 (Friday)
Time: 11:30am-12.00pm
Venue: 6th Floor Meeting room, SA Block, FES, UTAR, KL Campus, Jalan Genting Kelang, 53300 Kuala Lumpur
Speaker: Tat-Jun Chin
Title: Robust Fitting and Model Selection in Computer Vision
Abstract: One of the earliest examples of robust model fitting in Computer Vision is to detect lines and curves in images using the Hough Transform (Duda and Hart, 1972). Since then robust model fitting techniques have become part of the standard toolbox for Computer Vision researchers and practitioners, especially on geometry estimation problems where approaches like RANSAC dominate. Robustness is crucial in many cases as the data acquisition and preprocessing pipeline is inevitably imperfect, and this introduces outliers and noise into the data. A second important aspect is model selection, especially in the case of multiple structures, i.e. how many lines do we have in the image? Traditional model selection techniques dictate that the best model is the one that balances goodness of fit and model complexity. In this talk I will discuss ongoing work at The University of Adelaide to find new paradigms for robust model fitting and model selection. In particular I will report recently proposed methods that cast robust model fitting and model selection as statistical learning problems. These new methods offer the promise of more robustness and applicability to a wider range of problems.
Speaker Bio:
Dr Tat-Jun Chin is a Lecturer in the School of Computer Science at The University of Adelaide. He received his Ph.D. in Computer Systems Engineering from Monash University, and B.Eng in Mechatronics from the Universiti Teknologi Malaysia (UTM). Dr Chin's current research focusses on robust model fitting approaches for Computer Vision applications. In particular he is interested in searching for new paradigms for the fundamental underpinning of robust model fitting approaches, and to extend the applicability of robust model fitting approaches to a wider range of problems and operating conditions.

The seminar is open to all staff, students and public. Admission is free and no registration required. Please encourage your students / research counterparts to attend this seminar.

CCIS Seminar: Multimedia Forensics & Inverse Light Transport: Introduction and Open Problems

2011-09-27T00:54:00.002+08:00

Dear Colleagues,
Centre for Computing and Intelligent Systems (CCIS), Universiti Tunku Abdul Rahman (UTAR) will be organizing the following open seminar:

Date: 7 Oct 2011 (Friday)
Time: 11:00am-11.30am
Venue: 6th Floor Meeting room, SA Block, FES, UTAR, KL Campus, Jalan Genting Kelang, 53300 Kuala Lumpur
Speaker: Tian-Tsong Ng
Abstract: Multimedia forensics arises from a question as old as photography: When is seeing is believing. Today, images and videos are as commonly exchanged and shared as are text messages when Internet was first invented. Visual content has become inseparable from our daily communications and is increasingly used as personal records or evidence in courts of law. I will introduce multimedia forensics and describe the open problems awaiting answers.
I will also briefly talk about a different research topic in computer vision and graphics known as inverse light transport. Light transport is function that maps the light sources and the images. Inverse light transport can be considered a general framework for tackling low-level computer vision problems and the research is still at its infancy. I will introduce the related early works and the open problems.
Speaker Bio:
Tian-Tsong Ng is a research scientist in Institute for Infocomm Research, Singapore. He received his B.Eng in Electrical Engineering from University of Malaya in 1998, his M.Phil. in Signal Processing from Cambridge University in 2001 and his Ph.D. in Electrical Engineering from Columbia University in 2007. His research interest lies in computer vision, computer graphics and image forensics.
He won the Microsoft Best Student Paper Award at ACM Multimedia Conference in 2005, the John Wiley & Sons Best Paper Award at the first IEEE Workshop in Information Security and Forensics in 2009 and I2R Best Paper Award in 2010. He is a Commonwealth Scholar and an A*STAR Overseas Graduate Scholar.

The seminar is open to all staff, students and public. Admission is free and no registration required. Please encourage your students / research counterparts to attend this seminar.

The 16th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2012)

2011-09-20T18:04:00.002+08:00

PAKDD 2012: The 16th Pacific-Asia Conference on Knowledge Discovery and Data Mining
(Proceedings of the conference will be published in LNCS by Springer).

Kuala Lumpur, Malaysia, May 29 - June 1, 2012
Conference web site: http://pakdd2012.pakdd.org/

Important Dates Paper Submission Due: 2 October, 2011 (Sunday)
Author Notification: 30 December, 2011 (Friday)
Camera Ready Due: 22 January, 2012 (Sunday)

Conference Scope
The Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD) is a leading international conference in the areas of data mining and knowledge discovery (KDD). It provides an international forum for researchers and industry practitioners to share their new ideas, original research results and practical development experiences from all KDD related areas, including data mining, data warehousing, machine learning, aritificial intelligence, databases, statistics, knowledge engineering, visualization, and decision-making systems. The conference calls for research papers reporting original investigation results and industrial papers reporting real data mining applications and system development experience.

Topics
The topics of relevance for the conference papers include but not limited to the following:

Novel models and algorithms
Clustering
Classification
Ranking
Association analysis
Anomaly detection
Data pre-processing
Feature extraction and selection
Mining heterogeneous data
Mining multi-source data
Mining sequential data
Mining spatial and temporal data
Mining unstructured and semi-structured data
Mining graph and network data
Parallel, distributed, and high performance data mining on the cloud platform
Privacy preserving data mining
Mining high dimensional data
Mining uncertain data
Mining imbalanced data
Mining dynamic/streaming data
Statistical methods for data mining
Visual data mining
Interactive and online mining
Mining behavioral data
Mining multimedia data
Mining scientific databases
Ubiquitous knowledge discovery
Agent-based data mining
Mining social networks
Financial data mining
Fraud and risk analysis
Security and intrusion detection
Opinion mining and sentiment analysis
Post-processing including quality assessment and validation
Integration of data warehousing, OLAP and data mining
Human, domain, organizational and social factors in data mining
Applications to healthcare, bioinformatics, computational chemistry,
Eco-informatics, marketing, online gaming, etc

All paper submissions will be handled electronically. Detailed instructions are provided on the conference home page.

Work Completion Seminar: Cheng Kam Ching

2011-09-19T19:07:00.001+08:00

Your are cordially invited to the following work completion seminar:

Date: 30 September 2011 (Friday)

Time: 2:30pm-3:30pm

Venue: ~~5A meeting room, 5th floor, SA block,~~ SE202, Setapak Campus, FES, UTAR, Kuala Lumpur. [map]

Admission: Free

Speaker: Cheng Kam Ching

Title: Development of Job Matching Algorithm with Collective Learning Methods
Abstract:

Hiring or job match placing the wrong job seekers to the particular job posted by the employers is a costly mistake. As it is the key to keep the job site remain competitive. The classical online recruiting applications use only simple Boolean operations comparing the basic requirement information to generate the job matched results to job seekers. Therefore, it's often claimed to be irrelevant job matched or too many "hits" [1]. Job seekers must browse through a long list of job advertisements in a given query [2]. The problem of the classical method is about the data modelling. Job seekers need to fill up massive form-based basic requirement information. The data given is too subjective to define its importance. Hence, the matching may not be good. Besides, online recruiting application must be able to discover interesting jobs for recommendation. Job seekers are more likely excited to see a range of different job types in which really match his or her working profiles and own interests. This is the area addressed by this research described in this thesis. An investigation and development of a different method of job matching algorithm was undertaken. The primary objective of this research is to offer an intelligence job matching mechanism incorporated the relevant feedback approach. Real data (job title, job descriptions and job requirements of 3000 job advertisements) from Jobstreet will be pursued with detailed analysis for the validation of the proposed method.

Google DevFest 2011 @ Kuala Lumpur

2011-09-16T13:34:00.002+08:00

Google DevFest is an annual event for developers to learn about Google technologies and products. Since 2008, it has attracted thousands of app developers around the globe to participate in this event.

This year, Google will bring this popular DevFest event to Asia. Event at Kuala Lumpur here:

Date: 29 September 2011
Time: 9:00am - 7:00pm
Venue: UCTI College, Technology Park Malaysia, Bukit Jalil, Kuala Lumpur. [map]

	Track 1	Track 2
09:00 - 09:30	Keynote
09:30 - 10:30	Android Market for Developers (Tony Chan)
10:30 - 11:00	AM Break
11:00 - 12:00	Bleeding Edge HTML5 (Eric Bidelman)
12:00 - 13:00	Implementing Google Analytics (Vinoaj Vijeyakumaar)
13:00 - 14:00	Lunch
14:00 - 15:00	What's New and Important in Android (Tony Chan)	Displaying Large Geographic Datasets: Google Fusion Tables (Luke Mahe)
15:00 - 16:00	These Aren't the Site You're Looking For: Modern HTML5 Web Apps(Eric Bidelman)	Working Off the Grid: HTML5 Offline (Ernest Delgado)
16:00 - 16:30	PM Break
16:30 - 17:30	Insights to Drive Your Business Forward (Analytics Codelab) (Vinoaj Vijeyakumaar)	Fusion Tables (Geo Codelab) (Luke Mahe)
17:30 - 18:30	Demos
18:30 - 19:00	Q & A

Please register yourself here: https://devregsystem.appspot.com/?eid=234804

update of the event, please check out at the Google site here: http://code.google.com/events/devfests/2011/seasia.html#kualalumpur

CCSN Seminar by Dr. Jonathan Loo.

2011-09-14T14:25:00.001+08:00

Dear all,

The Center for Communication Systems and Networking will be organizing a university-wide seminar, as part of Brain Gain Malaysia Lecture Series, on the 21st of September from 11 AM to 12 Noon. The details of the seminar is given below.

Topic: Future Trends in Video Coding R&D and Security Provision of Internet of Things/6LoWPAN
Date: 21 September 2011
Time: 11 AM to 12 Noon
Venue: 6th Floor Meeting Room, FES, UTAR KL Campus, Setapak, Kuala Lumpur.
Speaker: Dr. Jonathan Loo, Reader, Middlesex University, United Kingdom

Kok Keong Loo a.k.a Jonathan Loo received his MSc degree in Electronics (with Distinction) from University of Hertfordshire, UK in 1998 and the PhD degree in Electronics and Communications from the same university in 2003. Between August 2003 and May 2010, he was a Lecturer in Multimedia Communications at the School of Engineering and Design, Brunel University, UK. Currently, he is a Reader (Associate Professor) at the School of Engineering and Information Sciences, Middlesex University, UK. At Brunel, he was a Course Director for the MSc in Digital Signal Processing programme. He also involved in several research projects where he was a Principal Investigator for a joint research project between BBC and Brunel University on Dirac video codec research and development, and a co-investigator for several European funded projects namely IST FP6 PLUTO IST FP6 ENABLE, and recently the IST FP7 3D-Vivant. So far, he has successfully supervised 10 PhDs as the Director of Studies. His research interests are in the general area of multimedia communications include visual media processing, video coding and transmission, wireless communications, digital signal processing, embedded systems and wireless network, protocols and security. In the past 7 years, he published in total of 128 publications (book chapters, journals, conferences) in the aforementioned areas. Recently, he has completed an edited book called "Mobile Ad Hoc Networks: Current Status and Future Trends" which will be published by CRC Press.

PhD positions in Computer Science and Engineering at IMT Lucca

2011-09-14T07:41:00.002+08:00

The institute for advanced studies IMT Lucca (Italy) announces 12 PhD positions in Computer Science and Engineering. The deadline for applications is September 28, 2011.

IMT (http://www.imtlucca.it/) is a research institute located in Lucca (Italy); courses are taught exclusively in English.

The PhD Program (https://www.imtlucca.it/phd_programs/computer_science_engineering/) coordinated by Rocco De Nicola aims at preparing researchers and professionals with broad training in the foundations of informatics as well as in applications to a variety of cutting-edge systems and disciplines.

A number of PhD students will be selected for working within the the two newly founded Research Unit:

SysMA (http://sysma.lab.imtlucca.it/) lead by Rocco De Nicola doing research on concurrent (distributed, mobile, autonomic) systems modelling and analysis and Dysco (http://dysco.lab.imtlucca.it/) lead by Alberto Bemporad doing research on control and optimization technologies.

In addition, students will be selected to work on topics related to medical imaging, and imaging for the life sciences http://users.eecs.northwestern.edu/~stsaft/#Openings

We hope that you may consider applying for and/or signaling these opportunities to colleagues and collaborators.

======= DETAILS ON THE ADVERTISED POSITIONS =======
6 IMT scholarship (12.423 EUR after taxes) plus accommodation and free meals (lunch and dinner).
1 MIUR scholarship (12.423 EUR after taxes) plus free meals (lunch and dinner).
5 positions to be funded with internal projects or third-party scholarships (negotiable salary) that come with a research budget of 3.000 EUR offered by IMT and free meals (lunch and dinner).

For further information please visit
http://www.imtlucca.it/phd_programs/call_for_applications/index.php
and/or contact the sender of this mail.

Stanford Machine Learning Class

2011-09-03T17:12:00.000+08:00

Similar to the Stanford University's AI Class, Prof. Andrew Ng, Director of Stanford's AI Lab, is also offering Machine Learning Class to the public. Sign up the class here: http://www.ml-class.com/

Below is the course description:
This course provides a broad introduction to machine learning, datamining, and statistical pattern recognition. Topics include:
(i) Supervised learning (parametric/non-parametric algorithms, support vector machines, kernels, neural networks).
(ii) Unsupervised learning (clustering, dimensionality reduction, recommender systems, deep learning).
(iii) Best practices in machine learning (bias/variance theory; innovation process in machine learning and AI).
(iv) Reinforcement learning.
The course will also draw from numerous case studies and applications, so that you'll also learn how to apply learning algorithms to building smart robots (perception, control), text understanding (web search, anti-spam), computer vision, medical informatics, audio, database mining, and other areas.

Ultra low cost computer from Raspberry Pi Foundation

2011-08-28T21:54:00.001+08:00

Raspberry Pi Foundation is set to launch its USD25 (Approx. RM75) ARM-based single-board computer by end of this year. It is just the size of a credit card. Plug in monitor and keyboard via HDMI and USB connection, respective, and voila!

* 700MHz ARM11
* 128MB or 256MB of SDRAM
* OpenGL ES 2.0
* 1080p30 H.264 high-profile decode
* Composite and HDMI video output
* USB 2.0
* SD/MMC/SDIO memory card slot
* General-purpose I/O
* Optional integrated 2-port USB hub and 10/100 Ethernet controller
* Open software (Ubuntu, Iceweasel, KOffice, Python)

The following video shows the prototype runs Quake 3 game:

Postgraduate Seminar at UTAR KL Campus

2011-08-25T10:54:00.001+08:00

Dear All,

We wish to inform that the Postgraduate Seminar for proposal defence
is scheduled as follows:-

Speaker: Chai Yung Joon
Date: 26/08/2011
Time: 10:00am-10:45am
Venue: 5A meeting room, SA Block, FES, UTAR KL Campus.
Title: Object Removal on Video Surveillance System
Abstract:
Nowadays, thievery is a common crime in the world. It can happen
anytime, from morning to night and anywhere from housing area to
public places. It makes the people feel so unsafe when they cannot
look after their valuable items such as travelling, out of home and so
on. Therefore, surveillance systems are employed at many places to
reduce the chance of the thievery happen. But normally be surveillance
systems need to be monitored by single operator. When human explore to
this kind of work for long time, they will be tire and easy make
mistake when then probability to miss dangerous situation will
increase. Thus, a semiautomatic surveillance system can be used to
help operator to detect thievery crime by giving some intelligences to
the system.
The major problem of surveillance systems is they normally only can
record the video and some of them able to detect there have something
moving in the scene. But the systems lack of the ability to detect and
track interest objects and human in video scene. When human and
objects can be tracked then further analysis can be done more
efficiently. Besides this, the systems also lack of the ability to
determine the person that take away the interested object. If the
system cannot give alert immediately, there will be a chance that the
thief can escape from scene.
This proposed system aims to address the above issues, which included
detect all interested objects in the scene at beginning of the video,
detect and track human in scene, and define the relationship between
human and interested objects in term of remove the object and occlude
the object. Besides this, the system aimed to run in 24 hours per day
which met the requirement of a standard surveillance system.
The proposed system can be divided to three modules, which are
foreground detection, human tracking and object removal detection. In
the scene, there will be a lot of interested objects and uninterested
objects. Foreground detection module will extract interested
foreground objects from background when the video start. After the
interested object detection in first module, all the interested
objects will be keeping tracked to use in object removal module. The
human tracking module will keep track on human who go into the video
scene until they are out of the scene. In object removal module, when
someone takes the interested objects in scene, the event should be
detected and the system able to tell which person took the objects.
This module also needs detect someone only occludes an interested
object but he/ she does not take away the object.
In conclusion, the proposed system aimed to detect and give alert when
people trying steal valuable items. It can help improve security level
in our society and lead us to have a better harmony environment.

Speaker: Kee Hong Sheng
Date: 26/08/2011
Time: 10:45am-11:30am
Venue: 5A meeting room, SA Block, FES, UTAR KL Campus.
Title: Unattended Object Detection Base on Video Surveillance Camera
Abstract
Nowadays, the awareness of terrorist attack prevention are becoming a
concerning issue in Malaysia. Many countries are also implementing
various security measurements to prevent terrorist attack. Setting up
surveillance cameras in strategic areas is one of the latest effective
measurements employed to detect such activities, where there are
surveillance cameras already set up everywhere, for instance,
government offices, airport terminals, car parks, staircases.
However, in order to monitor all the video scenes captured by the
surveillance cameras, it requires considerable human resource and
time. This project choose to detect unattended object, the main reason
is in a scene of surveillance cameras, the unattended object stay
static mostly all the time, and it's very hard to draw the attention
of the security personnel.
Hence, this project attempts to develop a system that is able to
automate the detection of possible occurrence of unattended object,
and hence alert the security personnel to pay attention to the
detected event, eventually some adequate measures can be carry out, to
minimize the probability of unpleasant incident to happen.
To develop a system that is able to widely implement, the system must
not required high resolution video image, where currently most of the
surveillance camera are low resolution or even maybe only able to
capture greyscale video. Processing high resolution video image will
also require higher processing power, or the efficiency of the system
will be affected. Hence, implement high resolution video image will
increase the implementing cost by increasing camera cost and also
processing unit cost.
The project proposes a human-object interaction analysis system that
detects human-object interaction, and the detected human-object
interaction will be monitored in the system. If the human-object
distance is too far, after a period of time, the object will be
considered as unattended object. The unattended object will be marks
up by the system, and the object owner will also be keep tracked.
In this project, the system dealing with human and object in a video
sequence, in most of the time, human and object will occlude each
other. The system scope will cover until object partially occlusion,
where fully occlusion include more uncertainty.
The Seminar is opened to public. All are welcome to attend the Seminar.

Research Assistants Needed

2011-08-18T17:27:00.001+08:00

RESEARCH ASSISTANT (RA)

Number of Vacancies: 2

Available at

Faculty of Engineering & Science, Universiti Tunku Abdul Rahman (UTAR)

KL Campus, Setapak, Kuala Lumpur

Project Title
Computer vision algorithm and software development for the application of urban traffic monitoring.

Project Leader
Associate Professor Dr. Tay Yong Haur (tayyh@utar.edu.my),
Department of Internet Engineering and Computer Science,
Faculty of Engineering and Science (FES), UTAR KL Campus.

Objective
To involve in the software development and continuous algorithm improvement of a computer vision-based software product for urban traffic monitoring.

Scope of Work

•          To evaluate the performance of the existing algorithm;
•          To recommend improvement on the existing algorithm;
•          To develop and evaluate an improved computer vision algorithm;
•          To integrate the improved algorithm into the software package;
•          To perform software development to enhance the overall software package
•          To work closely with research engineers from the industry.

Job Description

•          Remuneration of RM2000-2500 per month for 12 months.
•          The RA must help the project leader in all works associated with the research.
•          The RA must work closely with the sponsored industry partner to complete the research project.
•          The RA may register for Master Degree by research program (M. EngSc) with UTAR.

Requirements

•          Passionate in computer programming, especially in C/C++/C#
•          Strong interest and/or has working experience in image processing / pattern recognition project / GPU
•          Good English written and verbal communication
•          Good analytical skills
•          Self-motivated, requires minimal supervision, resourceful and willing to learn.
•          Available to start working preferably on 1 October 2011.

If you are interested in the RA position advertised above, please email your CV directly to Dr. Tay Yong Haur (tayyh@utar.edu.my) as soon as possible. Feel free to write to Dr Tay for any further inquiry of the project.

Only short-listed candidates will be called for interview.

Biometric School 2012 (Winter Edition) -- annoucement

2011-08-05T22:13:00.001+08:00

Please open the online version of this message if the HTML is not correctly displayed.

(approval pending)

Biometric School 2012 (Winter Edition)

Permai Rainforest Resort, Kuching, Sarawak, Malaysia

9-13 January 2012

www.biometricschool.org

Just announced : Biometric School 2012 programme

View the full programme.

Topics include :

Fundamentals of fingerprint recognition, Prof. Davide Maltoni (Italy)
Touchless fingerprint recognition & Biometrics and privacy protection, Prof. Jaihie Kim (Korea)
Iris recognition, Prof. Tieniu Tan (China)
Face recognition, Dr. Simon Prince (UK)
Linear and advanced classifier design for biometric systems, Prof. Kar-ann Toh (Korea)
Fusion strategies and template security in biometric systems, Dr. Karthik Nandakumar (Singapore)
Advanced techniques for multimodal biometric fusion : quality, cohort and client-dependent strategies, Dr. Norman Poh (UK)
Biometric-key computation, Dr. Andrew Teoh (Korea)

Remember to book by 31 October 2011 to get the best rate.

Talk by Prof Brian A. Barsky, UC Berkeley

2011-08-05T10:40:00.000+08:00

You are cordially welcome to attend the talk by Prof Brian A. Barsky on various aspects of computer science teaching and research at University California Berkeley.

Date: Monday, 8th August 2011

Time: 1:00PM ~ 1:45PM

Venue: IDK5, Kampar Campus,Universiti Tunku Abdul Rahman

Speaker: Prof. Brian A. Barsky

Brian A. Barsky is Professor of Computer Science and Affiliate Professor of Optometry and Vision Science at the UC Berkeley. His research interests include computer aided geometric design and modeling, interactive three-dimensional computer graphics, visualization in

scientific computing, computer aided cornea modeling and visualization, medical imaging, and virtual environments for surgical simulation.

Abstract:

In his talk, Prof. Barsky will share some of his insights and experience as a senior academic staff on various aspects of computer science teaching and research at UC Berkeley, which has been consistently one of the top ranked computer science schools in the United States and globally. UC Berkeley staff and alumni have garnered a total of 15 Turing awards (which is considered to be the "Nobel Prize" of computing) since its inception. It has been the source of common computer science products such as Berkeley Unix, relational databases, RAID (Redundant Arrays of Inexpensive Disks), Reduced Instruction Set Computing (RISC) and Simulation Program with

Integrated Circuit Emphasis (SPICE).

Attending the Stanford University's AI course

2011-08-04T23:35:00.004+08:00

Missing your Artificial Intelligence class? ;) Register yourself to attend the Introduction to Artificial Intelligence course offered by Stanford University, together with the students in Stanford University, for free!

The course will be taught by Prof. Sebastian Thrun, and Peter Norvig, the co-author of the popular AI book, AI: A Modern Approach.

The course will run from October to December 2011. You will be expected to watch the same lectures, complete the same assignments, and take the same examinations as the Stanford's students who sign up for the course. You will receive grades for your assignments and examinations, and will get certificates created by the professors if you manage to complete the course!

Below is the schedule:

Date	Content	Homework & Exams
Week of Oct 3	Overview of AI, Search	Assignment 1 due Oct 9
Week of Oct 10	Statistics, Uncertainty, and Bayes networks	Assignment 2 due Oct 16
Week of Oct 17	Machine Learning	Assignment 3 due Oct 23
Week of Oct 24	Hidden Markov models and Bayes filters	Assignment 4 due Oct 30
Week of Oct 31	Markov Decision Porcesses and and Reinforcement Learning	Assignment 5 due Nov 7
Week of Nov 7	Adversarial planning (games) and belief space planning (POMDPs)	MIDTERM EXAM due Nov 13
Week of Nov 14	Logic and Logical Problem Solving	Assignment 6 due Nov 20
Week of Nov 21	Image Processing and Computer Vision	Assignment 7 due Nov 27
Week of Nov 28	Robotics and robot motion planning	Assignment 8 due Dec 4
Week of Dec 5	Natural Language Processing and Information Retrieval
Week of Dec 12		FINAL EXAM due Dec 18

If you are interested, please register at http://www.ai-class.com/ before 10 Sept 2011.

Watch the introductory video here:

Research Assistant-ship at FES, UTAR KL.

2011-06-24T09:52:00.001+08:00

RESEARCH ASSISTANT (RA)

One (1) Position

Available at Faculty of Engineering & Science,

Universiti Tunku Abdul Rahman (UTAR)

KL Campus, Setapak, Kuala Lumpur.

Project Title

Super Resolution of Endoscopy Images based on Machine Learning and Image Processing Techniques.

Project Leader

Associate Professor Dr. Tay Yong Haur (tayyh@utar.edu.my),

Department of Internet Engineering and Computer Science,

Faculty of Engineering and Science (FES), UTAR KL Campus.

Project Objectives

To develop algorithms to enhance the quality of the endoscopy images

Scope of Work

• Literature Study

• Data Collection

• Algorithm development

• System Testing and Evaluation

Job Description

• Remuneration of RM2000 per month for 12 months.

• The RA must help the project leader in all works associated with the research.

• The RA may register for Master Degree by Research program (M. EngSc) with UTAR.

Requirements

• Excellent C/C++/C# programming skills

• Excellent English written and verbal communication

• Excellent analytical skills

• Experience in image processing or machine learning techniques

• Self-motivated, requires minimal supervision, resourceful and willing to learn.

• Available to start working preferably on 1 July 2011.

If you are interested in the RA position advertised above, please email your CV directly to Dr. Tay Yong Haur (tayyh@utar.edu.my) as soon as possible.

Only short-listed candidates will be called for interview

CCIS Seminar: A New Horizon? Korea NEXT Growth Engine : A Personal Experience by Dr Lau Phooi Yee

2011-06-03T07:47:00.001+08:00

Dear all,

You are cordially invited to attend the following talk by Dr Lau Phooi Yee.

Date: 6 June 2011 (Monday)

Time: 1.55pm - 3pm

Venue: A117J, UTAR Perak Campus, Kampar.

Speaker: Dr Lau Phooi Yee

Title: A New Horizon? Korea NEXT Growth Engine : A Personal Experience

Synopsis:

In today's global economy, companies aggressively fund universities and research labs to bring in fundamental innovations. Innovation becomes a priority especially through government-industry-academia collaboration scheme as the global interconnected networks reach commercial maturity in late-1990s, when the Internet came into existence. This can only be accomplished through close-knit research and development programs. For example, we see mobile phone is being developed into a mobile Internet. But innovations do not stop here. It will continue to grow. Why? Because the future of mobile communications still requires many technologies to be developed to handle huge data volumes. The evolution of devices through easy to use interfaces, e.g., touchscreen, has generated huge data traffic which needs to be efficiently managed. Content services via digital network have emerged, such as cloud computing, and are likely to take over as the primary way of content consumption around the world. Just like water or electricity which is readily available when moving into a new home, the basic access to content will be bundled into access to digital networks. This is already happening with the development of key technologies. In the future, it is possible that mobile telecommunication technologies will become one of the biggest innovation surges in business and private communication. This talk discuss some well funded innovations and key technologies that reflect the technological sophistication from pre-Internet era to post-Internet era and provide an overview of future research and development efforts in Korea.

Invited Talk: Emerging Computing Trends and Their Impact on System on Chip Design - Dr. Ong Sze Wei

2011-03-29T23:42:00.001+08:00

Dear all,

The Department of Electronics and Electrical Engineering and the
Center for Communication Systems and Networking (CCSN) will be
organizing a seminar by Dr. Ong Sze Wei. The details are given below.
Faculty members are invited to attend the talk.

Title: Emerging Computing Trends and Their Impact on System on Chip Design
Speaker: Dr. Ong Sze Wei
Date: 4th April 2011
Time: 1:30 PM
Venue: 6th Floor Meeting Room, FES, UTAR KL Campus, Setapak, Kuala Lumpur.
Abstract:
The global demand for high-performance computing and mobile Internet
connectivity continues to escalate as the appetite for media contents
grows. This demand, combined with the growing emphasis on small form
factor, is driving the need for System-on-Chip (SoC) design. This talk
will focus on the technology drivers, characteristics and challenges
of SOC design and development. The presentation will also describe
Intel's strategy in delivering a new category of smart SoC designs as
well as the design teams which deliver them.
Biography:
Dr. Ong Sze Wei is senior microprocessor design manager of Intel
Penang Design Center (PDC). He joined Intel in 2002 and has led
multiple design automation domains of Intel Atom(tm), Pentium(r) and
Core(tm) microprocessor projects. He is also the program manager of
Intel Technical Development Work Group where he champions strategic
training and development initiatives for the design teams in Malaysia.
In this capacity, Dr. Ont founded Intel-MDEC Advanced VLSI
Train-The-Trainer Program for 18 universities in Malaysia. He also
co-founded Northern Corridor Analog and Digital Design Program, which
is a catalyst program for developing advanced analog and digital
design engineers for the northern region of Malaysia. Dr. Ong is the
industry advisor for Universiti Sains Malaysia, Universiti Tunku Abdul
Rahman and Universiti Tenaga Nasional. He received his bachelor's,
master's and doctorate degrees in Electrical and Computer Engineering
from the University of Tennessee, U.S.A.

With thanks and regards,

Dr. Morris Ezra,
Asst. Professor,
Faculty of Engineering and Science,
University Tunku Abdul Rahman,

Work Completion Seminar: Research and Development of a Learning-Based Vehicle License Plate Detection Algorithm

2011-03-17T00:02:00.001+08:00

Dear all,

You are most welcome to attend the Work Completion Seminar for Mr Ho Wing Teng (Master of Computer Science), on his research title of Research and Development of a Learning-Based Vehicle License Plate Detection Algorithm. Details are as follow:

Date: 21/03/2011 (Mon)

Time: 11:00am-12:00pm

Venue: 5A meeting room, 5th floor, SA Block, FES, UTAR KL Campus.
Research title: Research and Development of a Learning-Based Vehicle License Plate Detection Algorithm

Candidate: Ho Wing Teng

Abstract:

The objective of this project is to research and develop a learning-based license plate detection framework with Spatial Horizontal Edge Variation features. As the growing of vehicle usage and the importance of automation in license plate recognition, license plate detection is one of the essential modules in the automation system. There are many existing algorithms available in the market in finding the license plate; in this project we introduce a framework that will learn from the image dataset and generate a set of classifiers to detect the targeted object. We tested our proposed framework on Malaysian vehicle license plate detection, and further experiment on Text detection. Spatial Horizontal Edge Variation feature is a simplified edge feature that can extract the edge information of the object through the horizontal orientation to differentiate between license plate and non license plate. Our framework selects a set of best features from a large set of feature set and generates a linear tree classifier to perform the object detection. From our experiments and analysis, we found that vehicle license plates have similarity with text that consist several characters. The changes of the edge information from the left to right horizontally through the license plate can actually provide useful information to distinguish whether it is a license plate or otherwise. In our project, we did further testing on the framework for detecting text from natural scene and side view car detection. For our license plate detector, we tested with some of the Malaysian vehicle license plate and achieve recall rate of 0.93796 and precision rate of 0.55366.

--
regards,
Yong Haur

TAY Yong Haur,
Associate Professor,
Head of Programme (Master of Computer Science)
Department of Internet Engineering and Computer Science (DIECS),
Faculty of Engineering and Science (FES),
Chairperson, Centre for Computing and Intelligent Systems (CCIS),
Universiti Tunku Abdul Rahman (UTAR),
Kuala Lumpur Campus
Jalan Genting Kelang, 53300 Kuala Lumpur
Malaysia.
Tel: (603) 4107 9802
Fax: (603) 4107 9803
Email: tayyh@utar.edu.my or YongHaur.Tay@gmail.com
URL: http://staff.utar.edu.my/tayyh/
URL: http://www.utar.edu.my/

Public Lecture by Nobel Laureate Professor Douglas Osheroff

2011-03-10T15:14:00.001+08:00

Title: How Sciences Changes Our Life
Speaker: Professor Douglas Osheroff, Nobel Prize in Physics 1996
Date: 15 March 2011
Time: 9.30am
Venue: National Science Centre, Bukit Kiara, Kuala Lumpur
Admission: Free
Closing Date of Registration: 10 March 2011
Tel: 03-2694 9898 (Akma/Rafiqah)
Email: akma@akademisains.gov.my / rafiqah@akademisains.gov.my

INVITED TALK (Tuesday): Video Surveillance with Robotics and Automatic Plants Classification by Dr. Paolo Remagnino from the Kingston University, UK

2011-03-08T00:16:00.001+08:00

With cooperation from ICT and Computational Science (ICTCS) research
cluster, we would like to invite you to a talk "Video Surveillance
with Robotics and Automatic Plants Classification" by Dr. Paolo
Remagnino from the Kingston University, UK.

The details of the public talks are as follows :

Title : Video Surveillance with Robotics and Automatic Plants Classification
Date : 15th March 2011 (Tuesday)
Venue : Lecture Hall 2, Annexe Building, Faculty of Computer Science &
Information Technology, Universiti Malaya
Time : 10.00am - 12.00pm

Admission is free. Walk-ins are also allowed.

ALL ARE INVITED, but please kindly register your attendance in advance
with cs.chan@um.edu.my.

Abstract of the talk
The Digital Imaging Research Centre will be introduced. Two projects
led by the speaker at Kingston University will be introduced. The
first one a three partner consortium funded by the US Department on
homeland security on augmented video surveillance with the use of
robotic platforms and the second a UK wide project with a consortium
of three sites aimed at the development of pattern recognition methods
for the automatic classification of plants through the analysis of
leaf characters. The latter is a an effort with the collaboration of
the Royal Botanic gardens at Kew.

Biography
Dr Paolo Remagnino (1993, PhD, IEEE and AAAI member) is a Reader in
the Faculty of Computing, Information Systems and Mathematics at
Kingston University and research associate with Royal Botanic gardens
at Kew. Dr Remagnino is a researcher in image and video understanding
and one of the main promoters of intelligent and autonomous
environments. Dr Remagnino is an editor member of the Expert Systems,
the Image and Vision Computing, the International Journal of Robotics
and Automation, Pattern Analysis and Application and the IEEE
Transactions of Automation, Science and Engineering. His research
interests include image processing, computer vision, machine learning,
artificial intelligence and robotics. Dr Remagnino has published a
numerous number of papers in scientific journals and conferences and
is the editor of 4 book collections, published with Springer, in the
field of video surveillance and ambient intelligence.

Thank you..

--
Chee Seng Chan PhD
Senior Lecturer

Department of Artificial Intelligence
Faculty of Comp. Sci. & Info. Tech.
University of Malaya
50603 Kuala Lumpur
Malaysia

cs.chan@um.edu.my
cs.chan@ieee.org

Talk on interdisciplinary research between medicine and engineering by Dr Makoto Kawashima

2011-02-28T23:06:00.001+08:00

Dear All,

Professor Dr. Makoto Kawashima from Chubu University, Japan will visit UTAR, FES campus and give an hour presentation on the introduction of the latest studies on interdisciplinary research between medicine and engineering.

Date of presentation: 15 March 2011(Tuesday).
Time : 2.00pm – 3.30pm
Venue: SA Block, 6th Floor, Meeting Room, UTAR K.L campus, Setapak.

You all are welcome to attend the presentation.

Thank you.

Best Regards,
Chuah Yea Dat
Lecturer,
Department of Mechatronics and BioMedical Engineering
Faculty of Engineering & Science
Tunku Abdul Rahman University
Kuala Lumpur Campus
Jalan Genting Kelang, Setapak 53300, Kuala Lumpur

Technical seminar: Human Machine Interfaces Design

2010-11-15T22:46:00.001+08:00

You are cordially invited to the following talk.

Title: Human Machine Interfaces Design

Venue: SA228, FES, UTAR KL Campus.

Time: 3-4pm, Nov 22, 2010

Abstract

The research frontier of human machine interaction is pushed forward by inventions and integrations of robotics and medical electronics. The body movements of desktop robots are used to enhance online conversations of instant messaging. They are sociable human machine interface that bridges the virtual world and humans in an intuitive and humane manner. ECG is a bio-signal interface between human body and computer. The appearance of highly integrated ECG frontends will soon make it wearable and ubiquitous like mobile phones. Besides serious medical application, ECG can be applied to instant messaging, emotion regulation, and traffic safety.

About Speaker

Dr Xing Shusong has been with the School of Software, Nankai University in Tianjin, China from 2005. He obtained the PhD degree from Nanyang technological University, Singapore. His research interests are in human-robot interaction, human machine interfacing devices, embedded system design, and visualization. He has published more than ten papers and patents. He is the supervisor of Eling Student Society in Nankai University. His students have been the winners of a series of innovation competitions including Microsoft ImageCup.

--
Best Regards,

Wang Xin, Ph.D.
Assistant Professor
Department of Mechatronics and BioMedical Engineering
Faculty of Engineering and Science
University of Tunku Abdul Rahman (UTAR)
P.O.Box 11384
50744 Kuala Lumpur
email: wangx@utar.edu.my

Free Workshop: Image Processing, Image Analysis and Real Time Imaging Workshop - 16 Dec 2010

2010-11-10T19:27:00.001+08:00

Image Processing, Image Analysis and Real Time Imaging Workshop

16 December 2010, Kuala Lumpur, Malaysia

IEEE Signal Processing Malaysia Chapter and UTM RAZAK School of Engineering and Advanced Technology, will jointly organize a Symposium on Image Processing, Image Analysis and Real Time Imaging Workshop, to be held on 16 December 2010 at UTM International Campus, Kuala Lumpur.

This free event is for all IEEE members and non-members with the aim of promoting IEEE Signal Processing Malaysia Chapter to Malaysian audience. This event is also aim in bringing the university-industry community together to share and discuss the latest trends image processing, image analysis and real time imaging technology. The topics of this workshop include but not limited to:

Image Acquisition and Display: imaging sensors and acquisition systems; sampling, scanning and quantization; color reproduction; image rendering; display and printing systems; evaluation of image quality.

Image and Video Processing: image filtering; enhancement and restoration; image and video segmentation; image registration; stereoscopic and 3‐D processing; video and image sequence processing; morphological processing.

Image and Video Analysis: low‐level feature extraction – color, texture and shape; high‐level semantics; pattern recognition and classification; motion detection and estimation; motion analysis and object tracking; object, event and scene recognition.

Storage and Retrieval: Image, audio and video databases; multidimensional indexing; content‐based retrieval; semantics retrieval; automatic image and video annotation.

Coding and Transmission: still image coding; video coding; audio and speech coding; stereoscopic and 3‐D coding; compression standards; image and video transmission over wireless networks.

Methodology and Algorithms: mathematical models; statistical and probabilistic models.

Abstract Submission
Prospective presenters are invited to submit an abstract. The selected presenters will receive a certificate of appreciation as an invited speaker.

Deadline for abstracts submission is 30th November 2010.

Please email your abstract to norliza@ic.utm.my

Participant Registration:
Please email or fax the following form to norliza@ic.utm.my (fax. No: 03-2693-4844)

MAD TechVentures 2010 Conference (8th – 9th November, 2010)

2010-11-03T23:48:00.001+08:00

Please go to http://www.madtechventures.com/ for more info

Free attendance with Promo Code: MAD2010

mYprize: a 4G Global Developer Challenge

2010-10-18T22:56:00.001+08:00

Dear MSC Malaysia Status Companies,MDeC would like to invite you to
participate in mYprize a 4G Global Developer Challenge.

Ideas will be evaluated by a panel of top judges from Head of New
Business Development of Google, Associate Director of MIT MediaLabs,
Partner of Crosslink Ventures, a Silicon Valley Venture Capitalist,
Senior Consultant of Light Reading & Datuk Badlisham Ghazali, CEO of
MDeC.

YTL Communications Sdn Bhd is launching the world's first Nationwide
Converged 4G Network. As a partner, MDeC would like to extend this
opportunity for you to be part of the YTL Communications' 4G
Innovation Network.

Submit your innovative ideas, applications, solutions or devices and
YTL's 4G Innovation Network collaborates with you to bring it into
reality & to the market place together.

Registration & Submission Closing Date: 30 November 2010

For more info, please visit:
www.myprize.my

The 9th International Conference on Cryptology And Network Security (CANS 2010): Call for Participation

2010-10-03T18:03:00.001+08:00

----------------------------------------------------
CALL FOR PARTICIPATION
The 9th International Conference on Cryptology And Network Security (CANS 2010)
12-14 December 2010
The Royale Bintang, Kuala Lumpur, Malaysia
----------------------------------------------------

The 9th International Conference on Cryptology And Network Security
(CANS 2010) will be held in Kuala Lumpur, Malaysia on December 12-14,
2010. It will be co-organized by Multimedia University, Malaysia and
Universiti Tunku Abdul Rahman (UTAR). The main goal of the conference
is to promote research on all aspects of network security, as well as
to build a bridge between research on cryptography and network
security. We therefore welcome scientific and academic papers that
focus on this multidisciplinary area. Previous CANS have been held in
Taipei (2001), San Francisco (2002), Miami (2003), Xiamen (2005),
Suzhou (2006), Singapore (2007), Hong Kong (2008) and Japan (2009).
You are invited to attend as a participant in the conference.

CONFERENCE WEBSITE:
For updated information regarding fees and registration as a
participant, please visit CANS 2010 conference website at:
http://fist.mmu.edu.my/cans2010/

FURTHER INQUIRIES:
For inquiries and further information regarding the conference, please
send an e-mail to: goibm@utar.edu.my,
cans2010-pc-chairs@dimacs.rutgers.edu

CCIS Seminar: Topics in CGIV by Prof Lee Byunk Gook, Dongseo Univesity, Korea

2010-07-17T11:22:00.000+08:00

You are cordially invited to attend the following talks by Prof Lee Byunk Gook, Professor from Division of Information Engineering, Dongseo University, Busan, South Korea:

Title: Topics in Computer Graphics, Image Processing and Computer Vision

Speaker: Prof Lee Byung Gook, Dongseo University, South Korea
Date: 22 July 2010
Time: 3pm - 4pm
Venue: SE105, UTAR KL Campus, Setapak, Kuala Lumpur.

Title: Topics in Computer Graphics, Image Processing and Computer Vision
Speaker: Prof Lee Byung Gook, Dongseo University, South Korea
Date: 23 July 2010
Time: 11am - 12pm
Venue: IDK5, UTAR Kampar Campus, Perak.

Talk on "Next Generation Biometric Systems" by Anil K. Jain

2010-06-16T15:49:00.002+08:00

Next Generation Biometric Systems

Anil K. Jain
Michigan State University
http://biometrics.cse.msu.edu

Date: 22 June, 2010
Time: 10.00am-11.15am
Venue: DKG31 Lecture Hall, School of Computer Sciences, USM

ABSTRACT
Prevailing methods of human identification rely mainly on credentials (identification documents and passwords/PIN). However, these methods are not able to meet the growing demands for stringent security in heterogeneous applications like identity management, international travel, government welfare programs, and access control. As a result, numerous person identification applications are adopting biometric technology that automatically recognizes people based on their anatomical or behavioural characteristics. Unlike credentials, biometric traits (e.g., fingerprint, face, and iris) cannot be lost, stolen, and easily forged; they are also supposed to be persistent and unique. The use of biometrics is not new; for over one hundred years, law enforcement and forensic agencies have successfully utilized fingerprints to apprehend criminals or identify victims. But, as biometric technology permeates our society, it faces new challenges. One such challenge is the need to design identification systems for ultra large scale civilian applications such as the Government of India's planned Unique ID project, which is projected to cover a population of over one billion citizens. The design and suitability of biometrics technology for person identification depends on the application requirements with respect to recognition accuracy, throughput, user acceptance, system security, robustness, and return on investment. Therefore, the next generation biometrics technology must overcome many research challenges such as the ability to deal with poor quality biometric data, interoperability, user privacy, system security, and information fusion. This talk will present our ongoing research work on latent fingerprint matching, models for facial aging, template protection, extraction and matching of soft biometric attributes, biometrics of twins, and matching sketches to the corresponding facial images in support of next generation biometrics systems.

BRIEF BIODATA
Anil K. Jain is a University Distinguished Professor in the Departments of Computer Science & Engineering, and Electrical & Computer Engineering at Michigan State University. He received a B.Tech. from IIT, Kanpur (1969) and Ph.D. from Ohio State University (1973). His research interests include pattern recognition, computer vision and biometric recognition. His articles on biometrics have appeared in Scientific American, Nature, IEEE Spectrum, Comm. ACM, IEEE Computer1,2, Proc. IEEE1,2, Encarta, Scholarpedia, and MIT Technology Review.
He has received a number of awards, including Guggenheim fellowship, Humboldt Research award, Fulbright fellowship, IEEE Computer Society Technical Achievement award (2003), W. Wallace McDowell award (2007), IAPR King-Sun Fu Prize (2008), and ICDM 2008 Research Contribution Award for contributions to pattern recognition and biometrics. He also received the best paper awards from the IEEE Trans. Neural Networks (1996) and the Pattern Recognition journal (1987, 1991, 2005). He served as the Editor-in-Chief of the IEEE Trans. Pattern Analysis and Machine Intelligence (1991-1994). He is a Fellow of the ACM, IEEE, AAAS, IAPR and SPIE.
Holder of six patents in the area of fingerprints (transferred to IBM in 1999), he is the author of several books: Handbook of Biometrics (2007), Handbook of Multibiometrics (2006), Handbook of Face Recognition (2005), Handbook of Fingerprint Recognition (2003) (received the PSP award from the Association of American Publishers), Markov Random Fields: Theory and Applications (1993), and Algorithms For Clustering Data (1988). ISI has designated him as a highly cited researcher (his h-index is 86). According to CiteSeer, his book, Algorithms for Clustering Data is ranked # 91 in the Most Cited Articles in Computer Science (over all times) and his paper "Data Clustering: A Review" (ACM Computing Surveys, 1999) is consistently ranked in the Top 10 Most Popular Magazine and Computing Survey Articles Downloaded.
He was a member of The National Academies committees on Whither Biometrics and Improvised Explosive Devices (IED) and a member of the Defense Science Board (2008-2010).

ORGANISED BY:
Pusat Pengajian Sains Komputer, USM
Pusat Pengajian Kejutereraan Elektrik & Elektronik, USM
Pelantar Penyelidikan Teknologi Maklumat & Komunikasi, USM

Participants are encouraged to send an email of confirmation of attendance to:
Mr Jeffiz Ezuer Shafii, email: jeffiz@notes.usm.my

Gesture-based computing on the cheap

2010-06-02T09:43:00.002+08:00

http://web.mit.edu/newsoffice/2010/gesture-computing-0520.html

The hardware for a new gesture-based computing system consists of nothing more than an ordinary webcam and a pair of brightly colored lycra gloves. Photo: Jason Dorfman/CSAIL

CMS Seminar Series 7 May 2010

2010-05-05T16:06:00.001+08:00

========================================================
Centre of Mathematical Sciences Seminar Series

Date: 7 May 2010 (Friday)
Time: 2 - 3 pm
Venue: Meeting Room, Block PD, UTAR PJ Campus

Speaker: Dr. Tan Aik Choon, University of Colorado

Title: Translational Bioinformatics: From Bytes to Bedside and Back

Abstract:

Cancer is driven by multiple genetic and epigenetic alterations, including changes in DNA sequence, aberrations of copy number, abnormalities of chromosomes, and modification of DNA methylation. Today's powerful and high-throughput technologies have enabled researchers: (1) to characterize these genomic abnormalities, which have largely been studied in isolation, within individual cancer genome; and (2) to survey these changes across large numbers of cancer genomes in great detail. Two major findings revealed from these studies: human tumors are heterogeneous and genomic alterations occur in specific pathways to drive cancer development and progression. These findings have major implication to oncology drug development and cancer treatment. In this talk, I will give an overview on how to develop and employ bioinformatics to integrate these massive data from diverse sources and translate them into meaningful clinical applications.

===================================================

Dear all,

This week CMS seminar is from Dr. Tan Aik Choon of University of Colorado.

Aik Choon Tan, Ph.D., is an Assistant Professor of Bioinformatics at the Division of Medical Oncology, Department of Medicine, University of Colorado Denver School of Medicine. He also has a secondary appointment at the Department of Biostatistics and Informatics, Colorado School of Public Health. He is a member of the University of Colorado Cancer Center, Developmental Therapeutics Program and a faculty member of the Head and Neck Cancer Research Program.

Aik Choon's research interest is to translate the basic research into clinical trials for personalized cancer treatment by 1) developing novel bioinformatics approaches to integrate and interpret high-throughput and complex biological data sources for biomarker discovery, and 2) devising new computational strategies to identify aberrant signaling and metabolic pathways in tumors. He is the Director of the Translational Bioinformatics and Cancer Systems Biology laboratory (The Tan Lab). He is also the coordinator of the bioinformatics and biostatistics group of investigators working in developing new algorithms and statistical methods to analyze next-generation sequencing data at the UCD School of Medicine.

Aik Choon graduated from the University of Technology Malaysia with First Class Degree in Bioprocess/Chemical Engineering (2000). He earned his Ph.D. in Computer Science/Bioinformatics from the University of Glasgow in 2004. Subsequently, he received his post-doctoral training at the Johns Hopkins University Whiting School of Engineering (2004 – 2007) and at the Sidney Kimmel Cancer Center at the Johns Hopkins University School of Medicine (2007 – 2009) before he joined the faculty of the University of Colorado in 2009.

He is an editorial board member of the Cancer Informatics journal, and he is a regular peer reviewer for bioinformatics related journals. He is a member of the International Society of Computational Biology (ISCB), the American Association for Cancer Research (AACR) and the American Society of Clinical Oncology (ASCO).

Thank you very much.

regards,
Kheng

----
Dr. Goh Yong Kheng
Assistant Professor / Deputy Dean
Faculty of Information & Communication Technology
University of Tunku Abdul Rahman
13, Jalan 13/6, 46200 Petaling Jaya, Selangor, MALAYSIA
Tel: 03-7955-1511 Ext 105

Writing and Presenting Scientific Papers by Prof. Tu Bao Ho

2010-04-26T22:58:00.004+08:00

1st MSC Malaysia R&D Workshop 2010

2010-04-26T16:10:00.000+08:00

1st MSC Malaysia R&D Workshop 2010

Tuesday, 4 May 2010

Venue: Junior Ballroom 1, NIKKO Hotel, Kuala Lumpur

Postgraduate Subjects to be offered in May 2010

2010-04-26T11:30:00.003+08:00

These are the postgraduate subjects (Master of Computer Science and Master of Information Systems) to be offered in the coming May 2010 trimester.

Research Methods in IT
Directed Readings
Internetworking With TCP/IP
Object-Oriented Database Systems
Evolutionary Algorithms and Applications
Enterprise Information Systems Architecture
IT Outsourcing
Software Project Framework
Information Systems for Market Intelligence

Augmented Reality Shooter Game

2010-04-24T14:32:00.000+08:00

EECS Graduate Study at Chiao Tung University, Taiwan

2010-04-19T11:52:00.001+08:00

Talk by Prof. Ivan Andonovic on "Evolution and Applications of Wireless Sensor Networks"

2010-03-23T11:04:00.001+08:00

Dear all,

You are cordially invited to the following talks:

Session 1:
Date: 24/3/10
Time: 11:00am - 12:30pm
Venue: UTAR PJ Campus PD206

Session 2:
Date: 25/3/10
Time: 2:00pm - 3:30pm
Venue: UTAR Perak Campus A009

Title: Evolution and Applications of Wireless Sensor Networks
Abstract: Advances in low cost wireless and processing hardware have stimulated a range of wireless sensor networks (WSNs) research and their subsequent deployment in a number of applications. Since cost is an essential factor in any development, their designs are driven according to issues such as restricted memory size, limited computation power and low power consumption. However advanced applications such as those demanded by the medical/health care, military and surveillance sectors require higher levels of network performance supporting time sensitive, guaranteed and secure high quality data delivery. These demanding applications drive the need to enhance first generation WSN concepts.

Speaker: Ivan Andonovic, University of Strathclyde, UK
Biography: Ivan Andonovic, Professor of Broadband Networks, graduated with a BSc (Hons.) in Electronic and Electrical Engineering in 1978 from the University of Strathclyde. He received a PhD in ‘Lithium Niobate Waveguide Devices’ in conjunction with neighbouring Glasgow University in 1982. He is presently the Head of the Centre for Intelligent Dynamic Communications Systems (CIDCOM) at the Department of Electronic and Electrical Engineering, Strathclyde University.
He joined Strathclyde in 1985 following several years at Barr and Stroud where he was responsible for the design, manufacture and test of optical guided wave devices. He has held a two year Royal Society Industrial Fellowship, in collaboration with British Telecommunications (BT) Labs during which he investigated novel approaches to broadband networking. Professor Andonovic’s research interests centre on the development of broadband networks, access and home networking and wireless sensor systems. He has edited two books and authored/co-authored six chapters in books and over 230 journal and conference papers. He has been chairman on the IET Professional Group E13, held a BT Short Term Fellowship, Visiting Scientist status at the Communications Research Laboratories of Japan, Visiting Professor at the City University of Hong Kong and Princeton University, USA. He is Topical Editor for the ‘IEEE Transactions on Communications’ and was Technical Programme Co-Chair for the recent ‘IEEE International Conference in Communications (ICC07)’. Prof. Andonovic is a Fellow of the IET, holds Senior Member status within the IEEE and is member of the Optical Society of America.
He was co-founder, Director and Chief Technology Officer of Kamelian Ltd., a high growth technology start-up focussing on the design and manufacture of advanced semi-conductor devices. He was also a member of flagship Scottish Enterprise (government agency for economic growth) establishment team of the Intermediary Technology Institutes (ITIs), aimed at bridging the gap between basic research and company growth. Presently he holds the ITI Techmedia Sponsored Chair in Broadband Networks facilitating the interaction between fundamental research in communications technologies and digital media and pre-competitive development of technologies that will form the input to commercialisation activities in Scotland.

Map to FICT-PJ

Skinput

2010-03-02T22:49:00.002+08:00

Skinput. Researchers from Carnegie Mellon University and Microsoft Research developed a prototype for user interface, called Skinput. The prototype allows users to use their own hands and arms as touch-screens by detecting the various ultralow-frequency sounds produced when tapping different parts of the skin.

Read here for more details.

Talk on "From NGN towards NGMN -Understanding the Relationship between 3GPP IMS and 3GPP Evolved Packet Core (EPC)" by Professor Dr. Thomas Magedanz

2010-02-25T19:27:00.004+08:00

Date: 11 March 2010 (Thursday)
Time: 8.30 am - 1.00 pm
Venue: Auditorium, MIMOS Berhad
Title: From NGN towards NGMN -Understanding the Relationship between 3GPP IMS and 3GPP Evolved Packet Core (EPC)

Speaker: Professor Dr. Thomas Magedanz

Kindly refer to the e-invitation as attached here and please send your RSVP to Ayu Hazlin at ayuhazlin@mdec.com.my or lina@mimos.my latest by 5 March 2010.

"Why Machines Need People" by Damon Horowitz, CEO of Aardvark

2010-02-13T07:22:00.002+08:00

SPICT-10 Website is Up

2010-02-12T15:08:00.000+08:00

Symposium on Progress in Information and Communication Technology 2010 (SPICT-10) will be held in Kuala Lumpur on 1-2 November 2010. The website is up and accessible here: http://spict.utar.edu.my/SPICT-10/index.html

--
regards,
Yong Haur

Talk by Prof Paul Roe on "Towards an Acoustic Environmental Observatory"

2010-01-30T14:42:00.003+08:00

Dear all,

You are cordially invited to the following talk:

Date: 2 February 2010 (Tuesday)
Time: 11:00 am - 12:00 am
Venue: PD206, FICT, UTAR Petaling Jaya Campus
Title: Towards an Acoustic Environmental Observatory
Speaker: Prof Paul Roe, Queensland University of Technology, Australia
Profile:
Professor Paul Roe received his MEng from the University of York in 1987 and his PhD from the University of Glasgow in 1991. He is currently a full professor in the Faculty of Science and Technology at QUT, in Brisbane Australia. At QUT he leads the Microsoft-QUT eResearch Centre, a collaboration between the Queensland State Government, Microsoft Research and QUT which is investigating smart tools for eResearch. Paul also chairs the QUT eResearch working party. He has published over 70 papers and received over $4M in competitive research funding. His research is focussed on smart tools which enable new forms and scales of research, particularly in the areas of distributed computing e.g. wireless sensors networks and end user programming languages.

Map to FICT-PJ

View Larger Map

Work Completion Seminar for Wong Chuan Ern (Master of Computer Science - A)

2010-01-30T14:36:00.001+08:00

Dear all,

You are invited for the Work Completion Seminar for Ms Wong Chuan Ern on next Monday. The details as as follow:

Date: 1 February 2010 (Monday)
Time: 3pm - 4pm
Venue: PD208A, FICT, UTAR Petaling Jaya campus.
Candidate: Ms Wong Chuan Ern
Thesis Title: Intruder Recognition Security System

--
regards,
Yong Haur

Standing Meeting Summary for December 2009 and January 2010

2010-01-29T17:48:00.001+08:00

Tou Jing Yi

He has attended the SPICT 2009 conference on 7th December and presented his paper on "Recent Trends in Texture Classification: A Review".
He has conducted experiment for verification-based recognition using covariance matrix. (Result: Accuracy of 98.33% on CAIRO dataset compared top 78.33% for GLCM as feature).

Hao Wooi Lim

He has been doing experiments on fast approximate nearest neighbor search. The 2 methods that were considered was Spill-Tree and KD-Tree. Spill-Tree and KD-Tree was the one provided by OpenCV 2.0. Spill-tree is about 9.7 times faster than KD-Tree on the MNIST dataset (when both are set to equivalent parameters). The lowest error achieved was 3.22%.
He has tried comparing Rob Hess's SIFT library with LIBSIFTFAST and the latter is roughly 9-10 times faster (tested on 640x480).
He has been doing experiment on how well Core Vector Machine (CVM) handles class-imbalance dataset as compared to LIBSVM. CVM seems to perform better by obtaining 94.7809% accuracy vs. for 87.4423% for LIBSVM).
He has made Core Vector Machine (CVM) to be able to train and test using multi-cores with OpenMP.
He has been experimenting on using CVM to train on MNIST. Current best result is obtained with the -c set to 10 and -e set to 0.00001, obtaining 1.48% error.
He has written a "warp generator" that is able to generate all possible projective transformation of a particular image. Unreasonable transformation (transformation that lead to significant loss of information) is discarded. What's left is around 1,846 possible transformation.
He has developed a generic natural text detector approach using MSER and SIFT-based unigram classifier. It is being applied to license plate detection.
He has manually labelled the Caltech Cars (Rear) data set (126 images) with labels of the license plate region.
He has prepared a real world license plate and artificially-generated characters dataset (630 train, 582 test) and tried training them with FANN and LIBSVM. The best result is obtained by SVM (86.9753%). The characters are scaled to 9x9 and binarized. Organizing the raw pixels row-wise or column-wise has no effect on the result.
He has modified his earlier algorithm for WoodCAS to be able to allow a cut to "cut through" a defect region.
He has attended the SPICT 2009 conference on 7th December and presented his paper on "Vehicle License Plate Detection Using Unigram Model and Difference-of-Surf Bigram Model with SVM".
He has been reading up on the following paper:

Linear Spatial Pyramid Matching using Sparse Coding for Image Classification

Lin Toh Koh

He has done some research in in OpenCV 2.0's HOG algorithm. HOG algorithm will first calculate the number of level scale (total scanning window) based on the image size with 1.05 scales. In each level scale, it doing the padding on the image size, getting the descriptor size (window size, block size, block stride, cell size, nbins and others) and the weight to detect human in particular level scale image.
He is not able to capture a full image for the CSIBot because it fails to get the exact fps in experiment coding but it getting the exact fps in pylon viewer. It might be the problem of library because it works before.
He has tested HOG project from Google code it given high false alarm.
He has succeeded to capture full image for CSIBot. The output images need to be enhance by capture more fps but it facing servo motor speed.
He is currently debugging opencv2.0, cvhog.cpp file. HOG computation, windows, block and cells size can be set for training and testing. He also looks at the OpenCV group and found that we are facing same problem of integration.
He is modifying the CSIbot board to slow down the servo motor by adding potential meter.
He will be continuing research on HOG algorithm.
He will be continuing capturing CSIbot image which has more fps to show more detail in a particular image.
He is reading a pedestrian detection project which is downloaded from google code. This project is mainly using HoG feature by Dalal with read.txt file. This project file is currently too messy with a lot of different directories where it spends time to tidy up. All the dataset are ready, and he will do training on different sizes of windows and cells, parameters when the program runs probably.
He will be continuing to do the training with the output result.
He will be continuing doing research on servo motor.
He has found out that the “black line” output image is causing by insufficient of light source. He has obtained some different output image with different focusing lens.
He is replacing the broken servo motor by the way tuning the parameter of fps with the new servo motor speed.
He is trying looking for other source to obtain the HOG features value. The output will be using for SVM training and testing.
He will be continuing setting up the CSIBot’s head to give the optimal output result.
He will be continuing research on methods of collection HOG feature value into text file.
He is building second CSIBot head.
He has reconstructed the CSIBot’s legs but it seems unstable.
He is testing on the line scan camera with the HF tech company and found that the problem of "black line" still occur. HF tech compay will give us support how to setting the line scan camera after confirm camera function well.
He is still working in the fundamental of HOG by reading the power slide and the thesis to get more stronger idea which propose by Dalal.
He will be continuing testing on the line scan camera and build it into CSIBot head possible make it functionality.
He will be continuing research on HOG algorithm and few methods are propose to enhance this algorithm including using cascade architecture and motion feature in a video.
He has attended the SPICT 2009 conference on 8th December and presented his paper on "Human Detection Using Motion Feature Extraction and Adaptive Boosting".
He is debugging into both algorithms where also trying understand the whole idea withe the Dalal's Phd Thesis (week 2). (Current understanding is author trying using the minimum size of 64 x 128 for each scanning window. It means that object size less than 64x128 might not be classified/detected).
He is collecting the CSIBot material (Servo, card board, basement roller, and others) to build the second CSIBot's body.
He will be continuing searching into both algorithms to find more detail steps of each step are running (week 2).
He will be building the second CSIBot's when the material are collected.

Chee Wei

He has finish modified the AdaBoost code on matlab to run the bigger dataset.
He will test on the previous dataset on the modified AdaBoost code.
He will working on the Ray features experiment. He will prepare the training and testing features vector from the ray features.
He is looking into the matlab AdaBoost code and modifying the feature selection part to handle the huge training file (Solving the issue of lacking runtime memory).
He is experimenting with the ray feature from the Kevin Smith.
He has attended the SPICT 2009 conference on 7th December and presented his paper on "Empirical Study of Boosted Weak Classifier in Object Detection Problem".

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.11.25

2010-01-29T14:59:00.000+08:00

Hao Wooi Lim

He has obtained the code for the simplified Convolutional Neural Network as described in the paper “Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis”. With some modifications, he obtained a test result of 1.15% on MNIST.
He has finish preparing the reports.
He has finish reading the paper on a dimensional reduction technique call “Stochastic Neighbor Embedding”.

Richard Ng

He will write the report for utarrf.
He will enhance on his thesis.

Lin Toh Koh

He has finished the proposal defence for his master.
He is testing on the Victor Adrian Prisacariu “fastHOG - a real-time GPU implementation of HOG” and it is success to run the technical parts. It give 1.7 second to finished calculate a 640 x 480 image.
He is also testing on OpenCV version 2.0 pedestrian detection found that it spend 250ms for 640x480 but its result is lower than Victor algorithm.
He is debugging into both algorithms where using HOG features and SVM as classification.
He will be continuing to debug into both algorithms to find more detail steps of each step are running.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.11.18

2010-01-29T14:56:00.000+08:00

Hao Wooi Lim

He has cleaned up and speeded up slightly Choon Hui’s CNN and fixed a couple of mistakes.
He has tested “Linear Spatial Pyramid Matching using Sparse Coding” method on Caltech-101 to obtain 73.5% accuracy. After removing an ambiguous category the accuracy jumps to ~75%.
He is finalizing the slides for proposal defense.
He is doing the code for visual dictionary generation (70% done).
He is running some experiments with Choon Hui’s CNN.
He will continue to work on the alternative simple method for WoodCAS.

Richard Ng

He has drafted the final report for utarrf.
He is enhancing on his thesis.

Lin Toh Koh

He is doing the practice and improvement the presentation slide.
He is doing the CUDA implementation using HOG. However library source is not provided by the author. He has sent the email to the author to gain more information.
He will be continuing to find out the coding problem which are facing now.

Lee Chee Wei

He is coding the Ray feature in the MATLAB.
He is currently focused on outlining the work completion.
He will complete the ray feature implementation.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Square for iPhone Demo by Kevin Rose

2010-01-27T08:24:00.000+08:00

CVT Seminar: Vehicle Guidance Technology by Dr Andrew Shacklock

2010-01-25T09:58:00.000+08:00

Date: 28th Jan 2010 (Thursday)

Time: 4pm - 5pm

Venue: DK5A, UTAR FES, KL.

Speaker: Dr Andrew Shacklock, Projective Space Ltd

Abstract:

Projective Space are a collaboration between practioners and researchers from the fields of robotics, computer vision and assistive technology. We are positioned to provide key navigation and interaction technology for the "Service Robotics" industry. To conduct research in autonomous and semi-autonomous robotic systems and implement the results in demonstrations and products for the defence, manufacturing and assistive technology industry sectors.

The current development of Projective Space are the latest innovation of using sensor fusion including mm wave radar in the vehicle following system. The image processing capability are essential for the success. We will share with you some of our experiences in color based texture in visual guidance applications.

regards,

Tan Ching Seong, Dr.

Chair

Centre for Vehicular Technology

email: tancs@utar.edu.my

Talk: Data Mining for Innovative Organisations: A Veritable Gold Mine!

2010-01-21T15:47:00.001+08:00

Date: Wednesday, 27 January 2010

Venue: Auditorium, Resource Centre Building, Technology Park Malaysia

PROGRAM

2:30 - 2:40 pm Welcome Address

2:40 - 3:30 pm

Data Mining for Innovative Organisations:
A Veritable Gold Mine!
Dr. Siew Eu-Gene
Senior Lecturer
Monash University Sunway Campus

3:30 - 4:00 pm

Innovation: A Way Forward to Business Sustainability!
En Azrai Shu’ib
Head of Incubation & Technopreneur
Development
Technology Park Malaysia

4:00 - 4:30 pm Q & A

4:30 pm Refreshments and Networking

Power Shutdown by TNB

2010-01-20T21:28:00.001+08:00

To all staff,

Please be informed that there will be a power shutdown by TNB on the following date and time to facilitate the maintenance work on their substation. Block PD and Block PE will be affected by this TNB exercise.

Date : 22 January 2010 , Friday
Time : 10.00 pm - 4.00 am
Area affected : Block PD & Block PE

Thank you for your attention.

Department of Estates and Facilities.

Jan 2010 Time Table for UTAR Master of Information Systems

2010-01-19T21:36:00.001+08:00

Monday 7pm-10pm Web Services by Too Chien Wen & Wong Lai Wan @ PD Lab6

Tuesday 7pm-10pm Strategic Management Information Systems byWong Siew Fan @ PD007

Wednesday 7pm-10pm Interactive Multimedia Learning by Siew Pei Hwa @ PA115

Thursday 7pm-10pm Pervasive & Intelligent Systems by Lee Yun Li @ PD007

Thursday 7pm-10pm Directed Readings by Tay Yong Haur @ PD008

Friday 7pm-10pm Research Methods in IT by Wong Siew Fan @ PD007

*14 study week from 18 January 2010 – 25 April 2010

Jan 2010 Time Table for UTAR Master of Computer Science

2010-01-19T21:27:00.002+08:00

Monday 7pm-10pm Web Services by Too Chien Wen & Wong Lai Wan @ PD Lab6

Tuesday 7pm-10pm Distributed Systems by James Ooi Ean Huat & Victor Tan Hock Kim @ PA115

Wednesday 7pm-10pm Computer Vision and Image Processing by Khor Siak Wang & Yap Shook Chin @ PA143

Thursday 7pm-10pm Pervasive & Intelligent Systems by Lee Yun Li @ PD007

Thursday 7pm-10pm Directed Readings by Tay Yong Haur @ PD008

Friday 7pm-10pm Research Methods in IT by Wong Siew Fan @ PD007

*14 study week from 18 January 2010 – 25 April 2010

Parrot AR.Drone: super-duper wifi-enabled, Linux-based quadricopter with vision

2010-01-19T07:04:00.001+08:00

SPICT-09 / WAIC-09/ 2nd R&D Forum

2010-01-16T10:59:00.002+08:00

Symposium on Progress in Information and Communication Technology (SPICT) 2009 was successfully organized on 7-8 December 2009, in conjunction with two other events, i.e. Post-International Conference on Neural Information Processing (ICONIP) Workshop on Intelligent Computing 2009 (WAIC-09), and 2nd Malaysia Development Corporation (MDeC) R&D Forum.

Despite the short notice given in our Call for Papers (CFP), SPICT-09 managed to received a total of 64 submissions, from over 31 different institutions in 12 different countries. The paper review was a double-blind peer review process, and each paper is sent and reviewed by at least 3 experts in the respective areas. Finally only 26 papers are accepted and registered to attend. The international response for this event was indeed very encouraging, and we hope that it will be an indicator of a future trend as this conference continues.

The two-day event run smoothly with high rate of presence from the participants and presenters. Out of the 26 presentation sessions, only 3 presenters are not able to present in the event. Some snapshots of the event are available here: http://spict.utar.edu.my/SPICTImageGallery09.html

To continue to organize an event which is affordable to the whole, and to encourage young researchers to meet and discuss, we have allocated a permanent server and URL for this purpose, i.e. http://spict.utar.edu.my. In this website, we have put in the proceedings of the conference here:

SPICT-09 proceeding http://spict.utar.edu.my/SPICT-09CD/index.html

WAIC-09 proceeding http://spict.utar.edu.my/WAIC-09CD/index.html

CCIS Seminar: Vehicle Routing in the Real World: A Case Study

2010-01-11T16:03:00.002+08:00

Dear colleagues,

Centre for Computing and Intelligent Systems (CCIS) is pleased to invite you for the following talk by Prof Sam Thangiah from Slippery Rock University, USA:

Date: 14 Jan 2010 (Thursday)
Time: 10.30am
Venue: PD208, UTAR FICT-PJ

Title: Vehicle Routing in the Real World: A Case Study
Abstract:
This presentation is on a school bus routing system implemented to solve a real-life multi‑depot, multi-destination, mix-fleet vehicle routing problem for a school district. The school district consisted of 1200 pickup points, 2000 students and 60 school buses. The precise distances between the pickup points were obtained using digitized road maps. The solutions obtained by the implemented system are compared with the routes used by the school district.

Speaker: Prof Sam Thangiah, Artificial Intelligence and Robotics Laboratory. Slippery Rock University
Profile:
Sam R. Thangiah has a bachelors and masters in computer science and a Ph.D. in computer science with emphasis in artificial intelligence and operations research. He initially worked on intelligent algorithms for routing cruise missiles in real time. He is currently the Professor and Director of the Artificial Intelligence and Robotics Lab. at Slippery Rock University and his research interests are in the routing and scheduling of robots/trucks for complex routing problems using intelligent algorithms.

Map to FICT-PJ

View Larger Map
--
regards,
Yong Haur

CMS Seminar: Fundamentals of Rational Complex Dynamics

2010-01-05T10:38:00.001+08:00

==============================================================================
Centre of Mathematical Sciences Seminar Series

Date: 12 January 2010 (Tuesday)
Time: 9 - 10 am
Venue: Meeting Room
Speaker: Dr. Liew How Hui, UTAR
Title: Fundamentals of Rational Complex Dynamics
Remark: This talk will be conducted in Chinese.

===============================================================================

Dear all,

I wish to apology that confusion caused for this week talk by Dr. Liew. I made a
typo mistake on the date of the talk. It is supposed to be 5 Jan (Tue) rather than
6 Jan (Tue). However, due to the confusion, many did not turn out. Dr. Liew was
kind to push the talk to next week Tuesday for the second part of the talk, but
had a discussion session with those came today on the examples of his first talk.
All are welcome for Dr. Liew's talk next week Tuesday 9-10 am.

Thank you very much.

regards,
Kheng

----
Dr. Goh Yong Kheng

Pranav Mistry: The thrilling potential of SixthSense technology

2010-01-04T21:31:00.001+08:00

At TEDIndia, Pranav Mistry demos several tools that help the physical world interact with the world of data -- including a deep look at his SixthSense device and a new, paradigm-shifting paper "laptop." In an onstage Q&A, Mistry says he'll open-source the software behind SixthSense, to open its possibilities to all.

Pattie Maes and Pranav Mistry demo SixthSense

2010-01-04T19:36:00.001+08:00

This demo -- from Pattie Maes' lab at MIT, spearheaded by Pranav Mistry -- was the buzz of TED. It's a wearable device with a projector that paves the way for profound interaction with our environment. Imagine "Minority Report" and then some.

MIMOS Research Seminar: Multi-camera Anomaly Detection

2009-12-16T11:48:00.001+08:00

WHAT : Multi-camera Anomaly Detection
WHO: Chen Change, LOY (Queens Mary College, University of London, U.K.)
WHEN: 1100 am – 1200 noon , Tuesday 22 December 2009
WHERE: MIMOS Berhad, Technology Park Malaysia, 57000 Kuala Lumpur

Admission : Free, but registration is encouraged..
Parking : There is ample free parking within the grounds of MIMOS Bhd.

Abstract :
We present a novel approach for detecting global behaviour anomalies in multiple disjoint cameras by learning time delayed dependencies between activities cross camera views. Specifically, we propose to model multi-camera activities using a Time Delayed Probabilistic Graphical Model (TD-PGM) with different nodes representing activities in different semantically decomposed regions from different camera views, and the directed links between nodes encoding causal relationships between the activities. A novel two-stage structure learning algorithm is formulated to learn globally optimised time-delayed dependencies. A new cumulative abnormality score is also introduced to replace the conventional log-likelihood score for gaining significantly more robust and reliable real-time anomaly detection. The effectiveness of the proposed approach is validated using a camera network installed at a busy underground station.

Profile :
Loy Chen Change is a PhD candidate in Queen Mary, University of London, U.K., under the supervision of Dr. Xiang and Prof. Gong. Previous to this, he was a researcher in Centre of Multimodel Signal Processing, Mimos Berhad. His work is focused on computer vision, pattern recognition and image processing. He has published in numerous prestigious computer vision conferences, e.g. ICCV, CVPR and BMVC.

All are welcome! However, as there are only a limited number of seats available, please come early to avoid the disappointment of not having a place to sit for the entire duration of the presentation.

We would appreciate receiving confirmation of participation (email: cs.chan@mimos.my or tel. 03 8995 5000 ext. 6270). See you there !

2nd MSC Malaysia R&D Exchange Forum 2009

2009-11-30T19:39:00.002+08:00

2nd MSC Malaysia R&D Exchange Forum 2009

Knowledge sharing in R&D on TelePresence and Video conferencing

Tuesday, 8 December, 2009

Venue: Royale Ballroom, The Royale Bintang Hotel, Kuala Lumpur

8.45 am Registration of Participants

9.10 am Welcoming Remarks

Ir. Prof. Dato' Dr. Chuah Hean Teik
President/CEO
Universiti Tunku Abdul Rahman (UTAR)

9.20 am Opening Remarks

Ms. Ng Wan Peng
Chief Operating Officer
Multimedia Development Corporation

9.30 am Panasonic HD Presence Communication System Development and Test Methodology

Mr. Eiichi Muramoto
Network Technology Leader
Immersive Communication Task Force, Corporate R&D
Panasonic Corporation Osaka

10.10 am ALM Technology for Multi-site HD Presence System and Standardization Activities

Dr.Ettikan Kandasamy
Principal Engineer
Panasonic R&D Centre Malaysia, Cyberjaya

10.50 am Tea Break

11.20 am Seamless AV Communication Utilizing Peer To Peer Technology & Challenges

Dr. Poo Kuan Hoong
Faculty of Information Technology
Multimedia University (MMU)

12.00 noon Panel Session

Moderator: Dr. Victor Tan (UTAR)
Panelist: Mr. Eiichi Muramoto
Dr. Ettikan Kandasamy
Dr. Poo Kuan Hoong

12.45 noon Networking Lunch

2.00 pm A TCP based High Definition Video Conferencing Platform.

Prof Dr.Sureswaran Ramadass
Director of the National Advanced IPv6 Centre of Excellence (NAV6),
IPv6 Domain Head for MYREN (Malaysian Research and Education Network)
Universiti Sains Malaysia (USM)

2.40pm Academic-Industrial Collaboration Forum: Industry academic collaboration: Experiences from ipVx-UTAR and NUS

Dr. Tommi Chen
CEO, ipVx Sdn Bhd.

3.20 pm Tea Break & Networking Session

3.50 p.m. Academic-Industrial Collaboration Forum: University – Industry Relationships - Bridging the Gap

Prof. Lee Sze Wei, Vice President (R&D and Commercialization),
Director of Institute of Postgraduate Studies & Research,
Universiti Tunku Abdul Rahman (UTAR)

4.30 pm Panel Session

Moderator: Dr. Liew Soung Yue (UTAR)
Panelist: Prof Dr.Sureswaran Ramadass
Dr. Tommi Chen
Prof. Lee Sze Wei

5.00 pm End of 2nd MSC Malaysia R&D Exchange Forum 2009

_________________________________________________________________

Dr. Ettikan Kandasamy Karuppiah

Principal Engineer

Panasonic R&D Centre Malaysia, Cyberjaya

Ettikan KK., with Ph.D (in the area of Distributed Computing), MSc (Master of Computer Science) and Bachelor of Computer Science, all from University Science Malaysia, obtained in the year 1997, 1999 and 2005 respectively. He is currently with Advanced Technology Development Group (ATDG), Panasonic R&D Centre Malaysia located in Cyberjaya, Malaysia as Principal Engineer (cum General Manager of R&D Group) being responsible in IP Multimedia communications protocol research activities. He is responsible for initiating collaborative research activities with Corporate Lab partners and formulating research process/methodology. Beside the team leading/managing effort, he also directly involved in R&D activities with 12 JP/International patents were successfully applied in the Home Networking area & Network Processing. Recently under his leadership ATDG/PRDCM was awarded STAR grant for strategic research activities at ATDG as part of Panasonic corporate research lab. Currently he is actively participating in embedded system communication related areas at ATDG.

Mr. Eiichi Muramoto

Network Technology Leader

Immersive Communication Task Force, Corporate R&D

Panasonic Corporation Osaka

Mr. Eiichi Muramoto received his Engineering Master degree from Information Science from Japan Advanced Institute of Science and Technology (JAIST). His research interest includes areas in Networking, Group communication, End to End QoS, Internet Testbed Application.

Dr. Poo Kuan Hoong

Lecturer, Faculty of Information Technology

Multimedia University

Dr. Poo Kuan Hoong received his Ph.D. degree from the Department of Computer
Science and Engineering in Nagoya Institute of Technology, Japan, and both his M.IT. and B.Sc. degrees from the Universiti Kebangsaan Malaysia, Malaysia.
His current research interest area spans peer-to-peer (P2P) networks and Distributed Systems. He is also interested in algorithm design and performance modeling of various distributed systems based on a variety of theories, including optimization, graph theory, and network coding. He is also interested in extensive measurement and statistical study of large-scale Internet systems, to explore problems and derive insights towards their improvement.

Dr. Tommi Chen

Chief Executive Officer

ipVx Sdn Bhd

Dr. Tommi Chen has been an early internet pioneer in Asia setting up the first ISP in the region in 1991. In Singapore he is best known as founder of Pacific Internet and is
credited as the ‘father’ of the internet for his work as a pioneer and catalyst for internet developments in Singapore. In 1995, he was invited by Telekom Malaysia to set up its ISP, TMNet. He was also once in research leading the R&D team for networking at the National University of Singapore and later becoming Deputy Director in charge of IT planning at the University.
Today Tommi is Chief Executive of ipVx Sdn Bhd. The company specializes in setting up ICT infrastructure and providing outsource services for data centres, internet service providers and large corporate clients. His forte is the internet, understanding its business implications and the service provider business. His strength is in his ability to create strong technological implementation capability and is known in the industry for his strategic insight in the internet service provider industry.

Prof Dr.Sureswaran Ramadass

Director of the National Advanced IPv6 Centre of Excellence (NAV6),

IPv6 Domain Head for MYREN (Malaysian Research and Education Network)

Universiti Sains Malaysia (USM)

Dr. Sureswaran obtained his BsEE/CE (Magna Cum Laude) and Masters in Electrical
and Computer Engineering from the University of Miami in 1987 and 1990 respectively. He obtained his PhD from Universiti Sains Malaysia (USM) in 2000.
Dr. Sureswaran's career included being a senior member of the technical staff of the
research team of MODCOMP, Florida. MODCOMP was a company focused on the
research, development and deployment of real-time operating systems. MODCOMP's customers included Jet Propultion Labs (JPL) and the National Aeronautics and Space Administration (NASA) for their remote space vehicles. Dr. Sureswaran later joined ICON Business System, Inc (Florida) as Senior Consultant in 1991. He was subsequently promoted to Vice President Engineering and was responsible for the entire engineering and R&D divisions of ICON.
Dr. Sureswaran's achievements include being awarded the Malaysian Innovation Award by the Prime Minister in 2007. Dr. Sureswaran was also the founder and headed the team that successfully took MLabs Systems Berhad, a high technology video conferencing company to a successful listing on the Malaysian Stock Exchange in 2005.
Dr. Sureswaran is currently one of the four steering committee members and the IPv6 Domain Head for MYREN (Malaysian Research and Education Network). He is also currently the Chairman of the Asia Pacific IPv6 Task Force (APV6TF). Dr. Sureswaran's is also a Primary Member of APAN as well as the Head of APAN Malaysia (Asia Pacific Advanced Networks).
Dr. Sureswaran is currently a Director of the Japanese AI3 project and is also the Head of the AI3 (Asian Internet Interconnections Initiative) for Malaysia. Dr. Sureswaran was also appointed as an advisor to the Japanese government's Ministry of Public Management, Home Affairs, Posts and Telecomm. (MPHPT).
Dr. Sureswaran has published over 70 research papers and 7 patents. He has written
chapters and provided materials for 4 books. He has also been a program committee
member for numerous international conferences and has reviewed papers for over 30 international and national level conferences.
Dr. Sureswaran has received over USD 8 million in research funding and grants from
numerous organizations in the 16 years that he has been with the university.

Prof. Dr. Lee Sze Wei

Vice President (R&D and Commercialization)

Director of Institute of Postgraduate Studies & Research

Universiti Tunku Abdul Rahman (UTAR)

Prof. Dr. Lee Sze Wei obtained his BEng (Hons) in Electronics and Optoelectronics,
MPhil., and PhD from University of Manchester Institute of Science and Technology, UK in 1995, 1996, and 1998 respectively. He specializes in telecommunication engineering and has been involved in various industry-university collaborative research projects and providing training programmes to the industry aimed at updating and upgrading the knowledge and skill of technologists and engineers in the industry.

Talk by Prof Sargur Srihari

2009-11-26T12:23:00.001+08:00

Weekly Standing Meeting 2009.11.11

2009-11-21T22:34:00.002+08:00

Hao Wooi Lim

He is doing the slides for proposal defence (40% done).
He is doing the code for visual dictionary generation (50% done).
He will continue to work on the alternative simple method for WoodCAS.

Richard Ng

He has done error analysis and the EER is 1.23%.
He will write the report for utarrf.
He will enhance on his thesis.

Lin Toh Koh

He has finished the proposal defence.
He has been reviewing the proposal by improving the technical and English sentence.
He has been doing the presentation slides for his proposal defence.
He will be doing pre-presentation so that to improve the presentation skills.
He will be continuing to do research on using SVM and CUDA.

Lee Chee Wei

He is coding the Ray feature in the MATLAB.
He is currently focused on outlining the work completion.
He will complete the ray feature implementation.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.11.04

2009-11-21T22:29:00.003+08:00

Hao Wooi Lim

He has finish revising the proposal.
He obtained a result of 30.9% accuracy on Caltech-50 (subset of the Caltech-101).
He is in coding a demo program for unsupervised objects discovery. Frame grabbing and keypoint detection module has been integrated.
He is preparing a slide for proposal defense.
He will continue to work on the alternative simple method for WoodCAS.

Richard Ng

He has examined the best bits code that contributes mostly to the recognition. The EER is 2.5%.
He is doing error analysis.
He will enhance on the iris recognition algorithm.

Kping Chin

He has successfully connected to camera and display the image in a window.
He is going to try capturing some image.
He will capture the image and store it as bmp file.

Sim Chee Choong

He is going to broadcast voice message to one client PC.

Lee Chee Wei

He has gone through the GML AdaBoost MATLAB toolbox. He has gone through the code and currently modifying to handle the huge training data input. He will be able to use the AdaBoost feature selection method once complete the modification.
He has running the training on the smaller set of data by bootstrapping method. He is now try to combine few classifier by additive method.
He is coding the Ray feature in the MATLAB.
He is currently focused on outlining the work completion
He will complete the ray feature implementation.

Lin Toh Koh

He has read through the “Max-Margin Additive Classifiers for Detection” and “Human Detection Using Partial Least Square Analysis”.
He has finished the first three chapters for proposal defense and the chapter of research methodologies is undergoing.
He has been writing the research methodologies chapter for proposal defense.
He has been testing on the technical part of using CUDA-based Pedestrian detection using Dalal and Triggs pedestrian detector. The author trying to explain the using others library such as blitz++, boost HOG also the CUDA SDK.
He will be continuing in writing the proposal defence which define clear of using SVM and CUDA. He also do the comparison between others machine learning and before and after using CUDA experiences.
He will be continuing testing on the technical part of using CUDA-based.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.10.28

2009-11-21T22:21:00.005+08:00

Hao Wooi Lim

He has done a survey on the current state of the art papers on keypoint detection and keypoint descriptor.
He is in coding a demo program for unsupervised objects discovery.
He is revising its proposal and slides for his proposal defence next month.
He will continue to work on the alternative simple method for WoodCAS.

Richard Ng

He has compared the correlation between rows to compensate for the deformation caused by pupil motion and imperfect segmentation. The EER is 2.64%.
He is examining the best bits code that contributes mostly to the recognition.
He will enhance the iris recognition algorithm.

Lee Chee Wei

He has run the testing on the OpenCV HOG-SVM detector (default classifier) on DC dataset (T1) with the result of TP: 80.7 and FP:16.2. (Error rate: 17.7%).
He has run the testing on the INRIA dataset with the OpenCV HOG-SVM detector. The test on the INTRIA dataset with TP: 52% and 11.43 FPFW.
He is implementing the Ray-feature. He is currently studying the feature and designs the implementation in coding.
He is building the classifier from the samples rejected by previous HOG-AdaBoost classifier for the combination of the multiple classifiers.
He will focus on his master work completion stage with outlining the required content.

Lin Toh Koh

He has combined the CSIbot coding into one file. This program will start with capturing the image for different angles. The image will be saved into bin format, convert to ppm format and lastly convert and save into jpg format.
He has practised the LIBSVM provided with the CJLin vehicle data set. (The feature value and testing value is provided).
He has been preparing to collect other features and convert them into LIBSVM format for training purpose. He would using haar-like or HOG feature in this case (3th week).
He has been testing on the CUDA SDK. He is looking some technical implementation method between CUDA and Adaboost for face detection.
He has been preparing the proposal and slide show with the two major things on the above (pedestrian detection and implementation CUDA) for the proposal defense.
He will be continuing in collecting the feature and convert into LIBSVM format for training and testing purpose.
He will be continuing searching technical information regard face detection Adaboost implementation using CUDA.
He will be continuing in enhancing the proposal report to make it more clear and specific to his master dissertation title.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Fwd: FICT Postgraduate Proposal Defense Seminar

2009-11-19T07:22:00.001+08:00

Dear all,

Please be informed that you are invited for the FICT Postgraduate Proposal Defense Seminar to be held on this Friday. All are welcome to the event.

The details are as follows:

Date: 20 Nov 2009 (Friday)
Time: PD206, FICT PJ.

No	From	To	Candidate Name	Programme	Title (Tentative)	Supervisor(s)
1	10:00:00 AM	10:30:00 AM	Hoh Ming Chee	MIS-A	The Development and Evaluation of an Independent Learner Courseware for Web Based Learning Environment (ILC-WBLE)	Dr Choo Wou Onn, Dr Siew Pei Hwa (Co-sup)
2	10:30:00 AM	11:30:00 AM	Ling Toh Koh	MCS-A	Pedestrian Detection Implementation Using CUDA	Dr Tay Yong Haur
3	11:30:00 AM	12:00:00 PM	Simon Lim Hao Wooi	MCS-A	Viewpoint-invariant Generic Object Recognition on static images via Temporal Semi-Online Semi-Supervised Learning	Dr Tay Yong Haur
4	12:00:00 PM	1:00:00 PM	Gan Ming Lee	PhD(ICT)	α+1: A New resource - Efficient path Protection Scheme	Dr Liew Soung Yue
5	1:00:00 PM	2:00:00 PM	Sor Kean Vee	PhD(ICT)	Parralel computation in RNA Structures Prediction	Dr Goh Yong Kheng
6	2:00:00 PM	2:30:00 PM	Lee Kelvin	MCS-A	Intruder Detection and Recognition through Behavioral Analysis	Dr Khor Siak Wang, Dr Lee Yun Li (Co-sup)
7	2:30:00 PM	3:00:00 PM	Ooi Zheng Lin	MCS-A	Detecting and Recognizing Intruder through Posture and Movement Activities	Dr Khor Siak Wang, Dr Lee Yun Li (Co-sup)

--
regards,
Yong Haur

IEEE I&M Workshop on Robotics and Instrumentation

2009-11-11T16:57:00.001+08:00

Dear Valued IEEE Members,

IEEE I&M Malaysia Chapter is organizing one-day workshop on Robotics and
Instrumentation - suitable for students, academicians and even hobbyists.
The workshop allows you to have hands-on experience on designing robots and
their instrumentation. Since, we have received many requests to postpone the
date due to examination period in many universities, therefore, the new date
has been selected and the details are as follows.

Date : Tuesday, 15 December 2009

Venue : Kulliyyah of Engineering, International
Islamic University

Time : 9.00am - 4.30pm

Price : RM50 (includes robotic kits, handouts,
certificates and refreshment)

Contact/ : A/Prof . Dr. Rini Akmeliawati

Registration by 09 December to <mailto:rakmelia@iiu.edu.my>
rakmelia@iiu.edu.my

Instructor : Dr. Riza Muhida

Itinerary :

09.00 Opening

09.15 Introduction to Robot: History, Structure and Components

10.30 Introduction to Microcontroller

11.30 Microcontroller Programming and data acquisition using Microcontroller

13.00 Lunch/Zhuhur Prayer

14.00 Programming For Robot

15.30 Design a Line Following Robot

16.00 Simulation of Line following Robot Competition

16.30 Closing

Short biography of instructor:

Dr Riza obtained his PhD in the field of Optical Electronics from Osaka
University, Japan. He obtained his Bachelor of Science in Physics from
Institute Technology Bandung, Indonesia; and his MSc in the field of Power
Electronics from Osaka University, Japan. He is currently an Assistant
Professor at the dept. Mechatronics Engineering, International Islamic
University Malaysia. His fields of interests: Instrumentation and
Measurement, Robotics, Sensors and Actuations. He has given many workshops
and trainings on Robotics and Instrumentation. He was one of the organizing
committee of the International Robot Olympiad 2008. He is an active member
of IEEE Instrumentation and Measurement Society Malaysia Chapter.

PS: We would appreciate if you can disseminate the information or the
attached poster to your colleagues/students/universities.

Thank you very much.

Sincerely,

Dr. Rini Akmeliawati

Chair of IEEE Instrumentation and Measurement

Malaysia Chapter.

Mimos Seminar invitation - Prof Dr Sergio Velastin

2009-11-09T11:44:00.001+08:00

To whom it may concern,

On behalf of Mimos Berhad, I have the honour to invite you and your colleagues to the presentation of Prof. Dr. Sergio Velastin, Kingston University, U.K., one of the leading expert in the field of digital imaging, smart notion and the analysis of human behavioural images in crowded situations, to take place on

Monday, 16th November 2009, 10:30a.m.
Mimos Berhad, Technology Park Malaysia, Bukit Jalil ( Seminar Room to be announced)

Prof. Dr. Sergio presentation will focus on "CCTV Video Analytics: Recent Advances and Limitations". Abstract of the presentation and a short biography of Prof. Dr. Sergio can be found below.

Abstract

The installation of closed circuit television (CCTV) cameras in urban environments is now commonplace and well-known. The UK leads the world with an estimated 4 million public cameras installed (20% of the world's deployment). Public attitudes to these systems reflect the balance needed between two conflicting requirements:

a) Concerns over invasion of privacy and fears of authoritarian control of the population. b) Welcoming the increased safety in public spaces and reductions in crime and antisocial behavior.

Recent events and what appears to be the effectiveness of the CCTV infrastructure in assisting law enforcers to understand how the events took place and the people involved, seem to have tipped the balance towards (b), at least momentarily. What we should not forget however, is that such events are thankfully very rare but that there is a cumulative significant effect of mundane daily events that we need to deal with. For example, it has been estimated that what in the UK is called "antisocial behavior" costs the country around € 5000 million a year. A single London Borough (municipality) spends annually around € 1 million to remove graffiti, 44% of women feel unsafe at bus stops at night, a single bus company in a major city is known to have replaced 8,000 windows in one year and a study showed that in a single day in the UK there were around 66,000 reports of nuisance or loutish behavior. At the same time, there are reports that when CCTV has been installed there has been a 35% reduction in crime over 5 years. Indeed, crime in general is decreasing while uncivil behavior is on the increase.

It turns out that one of the major limitations of conventional CCTV systems is the impracticality of deploying sufficient number of people to be in front of television screens observing largely uneventful video. As long as this is the case, CCTV will tend to remain a reactive tool. The inability of being truly pro-active, producing timely alarms and eventually being able to prevent incidents, is what ultimately limits these systems. There is a need to automate at least simple observation tasks through computer vision, a functionality that has become known popularly as "video analytics". The large size of CCTV systems and the requirement of high detection rates and low false alarms are major challenges., this talk will

illustrate some of the efforts that the research and industrial communities are making towards realizing automated means of detecting video events involving human activity. It will show the kind of progress made but also the current limitations of this technology.

Biography

Prof. Dr Sergio A. Velastin received the B.Sc. degree in Electronics, M.Sc. (Research) degree in Digital Image Processing and the Ph.D. from the University of Manchester Institute of Science and Technology (UMIST) in the UK, in 1978, 1979 and 1982, respectively.

Prof. Dr Sergio Velastin founded Ipsotek in 2001 and is widely considered a leading expert in the field of digital imaging, smart motion detection and the analysis of human behavioural images in crowded situations. Prof. Dr Velastin has now returned to academia where he is the Director of the Digital Imaging Research Centre (DIRC) at Kingston University. However, Prof. Dr. Velastin remains involved with Ipsotek in an advisory capacity. Prof. Dr. Velastin has published 2 books and over 150 papers refereed journal and conference. His research interests include computer vision for pedestrian monitoring and personal security as well as distributed visual surveillance systems.

Prof. Dr Sergio Velastin is a member of the IET, IEEE and the British Machine Vision Association (BMVA). He also served as a EPSRC grant reviewer and member of the International Union Public Transport (UIPT). Recently, Prof. Dr. Sergio has been awarded £11 million funding from the EU FP7 for a research project named "PROTECTRAIL".

The presentation will end approx. at 11.30a.m.

We would appreciate receiving confirmation of participation before 13.11.2009 (email: cs.chan@mimos.my or tel. 02 8995 5000 ext. 6270). Please be informed that light refreshment will be provided after the presentation. See you there !

Regards,

CS, CHAN

----------------------

Dr Chee Seng Chan

Senior Reseacher
Mimos Berhad
Technology Park Malaysia
57000 Kuala Lumpur
MALAYSIA
Email: cs.chan@mimos.my
Contact: +(60)3 8995 5000 ext. 6270

Weekly Standing Meeting 2009.10.20

2009-10-26T00:15:00.004+08:00

Hao Wooi Lim

He has obtained a result of 78.59% on Caltech-101 using Geometric Blur (GB), Histogram of Visual Words (PHOW), Self-similarity (SSIM). The best result I obtained is 81.3% by the 2007 paper “Image Classification using Random Forests and Ferns”.
He has obtained the skew spectrum features from our vehicle license plate dataset.
He has studied on Group theory basics for better understanding on the skew spectrum paper.
He is writing a program to train the Skew spectrum features obtained.
He will prepare a simple prototype for unsupervised objects discovery on a toy video.
He will continue to work on the alternative simple method for WoodCAS.

Richard Ng

He has optimized the parameter for correlation filter and the EER is 3.93%.
He is comparing the correlation between rows to compensate for the deformation caused by pupil motion and imperfect segmentation.
He will enhance on the iris recognition algorithm.

Lee Chee Wei

He is running the analysis for the result of the HOG-AdaBoost for DC dataset.
He is running the performance evaluation for the HOG-SVM in OpenCV for DC dataset.
He is implementing the Ray features for the Pedestrian-detection.
He will running the performance of evaluation on INRIA (and other) dataset on openCV HOG-SVM default classifier and HOG-AdaBoost classifier.
He will running the experiment on the bootstrapping with combining multi-classifier trained different bootstrapped sample space.

Lin Toh Koh

He has tested on the 2nd servo motor turning degree is consistent. This servo motor only turns maximum 160 to 170 degree where each turning point degree is less than 1 degree.
He has reviewed “A Parallel Implementation of 2D Wavelet Transform Using CUDA” paper. This paper briefly introduces the version of NVIDIA product, architecture (memory with different type NVIDIA’ model) and the programming. It also experiment on wavelet transform using CUDA for few different high resolution image.
He has been looking more information regards the parallel processing with CUDA. It pros and cons will be prepared for the proposal defences.
He will be continuing collect more information regard Compute Unified Device Architecture system and the implementation.
He will be discussing with Dr Hon to solve the basler camera too dark issue by sending back the camera to the vendor.
He will be continuing the research on the methods of training for SVM.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.10.14

2009-10-26T00:12:00.002+08:00

Tou Jing Yi

He has attended the UTAR Research Workshop in Kampar.
He is working on WoodCAS algorithm to randomly cut based on defects (Fixed previous problems due to 2 logical errors in the code. This algorithm can cut the wood into pieces that either obtain defects or did not obtain defects, an allocation algorithm can be applied onto the piece that does not contain defects for further cuts.)
He will be finalizing the basic cutting algorithm as a baseline.

Hao Wooi Lim

He has successfuly run 2 experiments on MNIST using hinton’s deep autoencoder and deep belief network. The Deep belief network obtain 0% error on training set and 1.09% error on testing set.
He has read up on skew spectrum of graphs features. It’s a method to convert a graph into a feature vector. The method is based heavily on spectral graph theory.
He has implemented skew spectrum of graphs features on vehicle license plate detection.
He is currently running an experiment on object recognition using Geometric blur, PHOW, Self-similarity features trained using a 1-versus-all Multi-Kernel SVM classifier.
He is implementing an alternative simple method for WoodCAS (Do allocation on the biggest possible bounding circle first, then do a greedy best fit algorithm to account for irregular shape and defects).
He will continue to work on the alternative simple method for WoodCAS.

Richard Ng

He has tuned the frequency of the correlation filter and the EER is 5.88%.
He is doing enhancement for the 1D correlation filter.
He will enhance on the iris recognition algorithm.

Lee Chee Wei

He has finish running the hog feature with f-score selection with 24k feature using Gentle Adaboost classifier.
He has read on the paper "Fast Ray Features for Learning Irregular Shapes ICCV2009" and "An HOG-LBP Human Detector with Partial Occlusion Handling ICCV 2009".
He will focus on the ray feature and implement the ray features in detection.
He will running the experiment of the hog and svm with the available dataset to find out the performance of the classifier by openCV.

Lin Toh Koh

He has tested on the basler camera image quality and it might be fault of the camera itself.
He has reviewed the cjlin paper with the topic of SVM with practical guide of training.
He has been looking for the basler camera vendor to solve the camera problem. It is possible for us to send back the basler camera if problem not solved.
He has been looking for the SVM practical and the detail of formula which simple to classify into different region by a straight line.
He will be discussing with Dr Hon to solve the basler camera too dark issue by sending back the camera to the vendor.
He will be continuing the research on the methods of training for SVM.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.10.07

2009-10-26T00:08:00.001+08:00

Tou Jing Yi

He has updated and submitted the camera-ready version for SPICT.
He has been exploring the wood cutting algorithm (The first idea is using histograms on x and y axis to identify areas big enough to fit in the target woodchips. The second idea is to use a primitive binary tree spanning method to cut down the wood log).
He will be developing the wood cutting algorithm based on primitive binary tree spanning.

Hao Wooi Lim

He has collected and organized various experimental codes implementing various AI/CV algorithms.
He has managed to run demos using code for generating GIST descriptor, PHOW, PHOG and various dimensionality reduction algorithms.
He is currently running an experiment on object recognition using Geometric blur, PHOW, Self-similarity features trained using a 1-versus-all Multi-Kernel SVM classifier.
He will continue on the 2nd part of the randomized cutting algorithm.
He will work on a simplified version of a prototype for unsupervised objects discovery on video for proposal defence.

Richard Ng

He has implemented the 1D correlation filter and the EER is 6.46%.
He is doing enhancement for the 1D correlation filter.
He will enhance on the iris recognition algorithm.

Lee Chee Wei

He has reading a paper title "Class-specific Hough Forest for Object Detection". He has download the demo code publish at the author website and exploring the code.
He has explore the openCV 2.0 and runing the HOG with SVM demo inside the OpenCV.
He has currently testing the result on the HOG featureas with 24k. (2nd week).
He will summarize the recent trend of the technique on Human Detection Problem.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

3rd MSC MALAYSIA R&D SERIES 2009 : Next Generation Wireless Network

2009-10-22T09:46:00.001+08:00

Dear all,

We are pleased to invite you for the event below. You may RSVP to thelagavathi@mdec.com.my. Looking forward to have you for the event. Thanks.

*Please feel free to forward this invitation to relevant contact.

Thanks & Regards,

Thelagavathi Atikan
Senior Executive - Technology Policy and Research
Capacity Development Division
Multimedia Development Corporation (MDeC)

DID: +603-8314 1842 |Fax: +603-8318 8786 |Email:
thelagavathi@mdec.com.my <mailto:khairulmazwan@mdec.com.my>

Talk on Research Grant, Journal Publication and Consultancy Work (22/10/09)

2009-10-14T21:53:00.001+08:00

Dear colleagues,

We are pleased to inform you that Faculty of Business and Finance
(FBF) will organize a half-day seminar on 22nd October 2009 (Thursday)
from 1.00p.m. - 5.00 p.m. The venue of the seminar is at Auditorium
A003, Blok A, UTAR Kampar Campus. Three speakers were invited to give
a talk on research grant, journal publication and consultancy works.
The details of the seminar are as follow:-

Topic 1: How to Publish Research Papers in the Reputable Journals Effectively?
Speaker: Dr. Alain Yee-Loong Chong, Associate Professor, INTI
University College, Malaysia

Topic 2: How to Apply External Research Grant?
Speaker: Dr. Chew Kok Wai, Senior Lecturer, Multimedia University, Malaysia.

Topic 3: How to Obtain Consultancy Works?
Speaker: Dr. Ch'ng Huck Khoon, Assistant Professor, Universiti Tunku
Abdul Rahman, Malaysia.

Attached are the resumes of the above speakers.

We look forward to welcoming you to the above talks. Please
disseminate this information to your staff and we would like to invite
all of you to come forward and join us in sharing the knowledge and
valuable information together.

If you would like to reserve a place please email or call our Deputy
Dean (R&D) Dr. Choong Chee Keong - choongck@utar.edu.my. We would be
very pleased to welcome you.

Thank you

Best regards,
Dr. Ooi Keng Boon
Dean and Associate Professor
Faculty of Business and Finance
Universiti Tunku Abdul Rahman
Perak Campus
Jalan Universiti
Bandar Barat
31900 Kampar, Perak, Malaysia
Tel: +605-4662323 Ext.1033
Fax: +605-4667407
Email: ooikb@utar.edu.my

Weekly Standing Meeting 2009.09.30

2009-10-01T10:53:00.003+08:00

Tou Jing Yi

He has communicated with Simon on the updates on WoodCAS project.
He has been enhancing the SPICT paper.
He has been helping Chee Wei and Toh Koh to proof read their SPICT papers.
He will be finalizing the SPICT paper for submission.
He will be assisting Chee Wei and Toh Koh to finalize their SPICT papers
He will looking into the WoodCAS algorithm.

Hao Wooi Lim

He has been playing and learning on LSH and ANN. The LSH library used is E2LSH.
He has ported E2LSH to compile under Visual Studio 2008.
He has compared the descriptor in OpenSURF with Rob Hess’s SIFT. The SIFT macthing demo in OpenSURF seems to be less rotation-invariant than the one in Rob Hess's SIFT matching demo program.
He has finish revising the SPICT 2009 paper based on the reviewes obtained
He has assisted Chee Wei to revise his SPICT 2009 paper.
He will work on block allocation algorithm for the WoodCAS.

Richard Ng

He has designed the OTSD filter and use three training images for each class and the EER is 19.07%.
He is implementing the 1D correlation filter to extract frequency information in the angular direction.
He will enhance on the iris recognition algorithm.

Lee Chee Wei

He has finish modified the conference paper according to the review of reviewer.
He has run the AdaBoost training with on the 20k HOG features with the F-score selection.
He is currently testing the performance of the F-score selected features from these larger set of HOG features.
He is compiling the result of the previous experiments and other similar experiment from other paper.
He will analyze the result of the performance by identify the wrongly classified samples.
He will focus on the feature selection method by enhancing with bootstrapping method.

Ling Toh Koh

He has synchronized the Pic 18 and Line scan camera by given .ppm output file.
He has been working on the output image which is given “black line” in every output pixel.
He has been working on capturing the multiple angles and append into one image file.
He will be continuing on testing and enhance the CSIBot on the output image and the multiple angles view.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.09.16

2009-10-01T10:45:00.005+08:00

Tou Jing Yi

He has edited and submitted for a conference.
He is enhancing SPICT ‘09 paper.

Hao Wooi Lim

He has finished implementing a random-cut algorithm for the WoodCAS project together.
He is currently working an enhancement of SIFT.
He will work on block allocation algorithm for the WoodCAS.
He will try to enhance the SPICT 2009 paper based on the reviews obtained.

Richard Ng

He has implemented the UMACE filter and its EER is 34.6%.
He is designing the MACE filter and the OTSD filter.
He will enhance on the iris recognition algorithm.

Chee Wei Lee

He has completed the training and testing on the HOG feature with the f-score.
He will focus on the enhancing the accuracy by focusing on the feature.
He will getting the result for the larger HOG feature portion to test on the same data set.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Hong Kong PhD Fellowship Scheme

2009-09-19T09:22:00.001+08:00

Dear all,

Please follow the link below if you are interested to pursue you PhD in HK.

http://www.ugc.edu.hk/eng/rgc/hkphd/hkphd.htm

thanks.

--
regards,
Yong Haur

Weekly Standing Meeting 2009.09.09

2009-09-15T15:06:00.004+08:00

Tou Jing Yi

He is currently doing some paper editing.
He will run more experiments for his next paper.
He will try to enhance the SPICT 2009 paper.

Hao Wooi Lim

He has played with the source code for a new improved version of SIFT called ASIFT that is completely invariant to affine transformation.
He is currently working on a simple prototype wood cutting optimization application.
He will continue to work on a draft algorithm for best wood cut search.
He will try to enhance the SPICT 2009 paper based on the reviews obtained.
He will work on a paper for ACIIDS 2010 on improving ASIFT with keypoints disambiguation.

Ho Wing Teng

He has revised on paper titled "Learning Informative Feature For Spatial Histogram-Based Object Detection", and modified the structure of the training program to Cascaded Histogram Matching. [p/s: previously applying the spatial histogram in Adaboost].
He has been testing on the spatial histogram (13th week).
He has been testing the modified training program(cascade Histogram Matching) with UIUC Car Test Samples).
He will be continuing on testing on the spatial histogram.

Richard Ng

He is implementing the correlation filter of iris matching.
He is designing the correlation filter that suits for iris recognition.
He will enhance on the iris recognition algorithm.

Chee Wei Lee

He has modified the f-score selection method to calculate the average error rate from the random selected samples ranging from the 0.01 to 0.09 threshold. He found the the best result come from the total of 1377 features.
He is currently testing the feature selected from f-score method to get the performance result.
He will modify the paper based on the comment from the reviewer.
He will test the result of the 27k feature selected from new f-score selection method.

Ryan Low

Linked PIC MicroController with Mini Cpu.
Linked Linescan camera with MiniCpu and get the raw buffer data.
Convert the raw data which is in binary form into text form.
Link all the Hardware to do finalization.

Ling Toh Koh

He has installed the libusb in Ubuntu Linux OS to control microcontroller PIC18. (It is able to set the speed and angle of the servo motor by burning the code into the PIC18.).
He has link up the line scan camera with PICO820 to record the image form which are success recorded in binary form.
He will be continuing build the CSIBOT into the PICO820 which might control the microcontroller and line scan camera at parallel.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.09.02

2009-09-15T14:54:00.003+08:00

Tou Jing Yi

He will enhance iconip 2009 paper.

Hao Wooi Lim

He has finished proposal 2 for Research Methodology.
He has finished proposal 2 presentation slides for Research Methodology.
He will continue to work on a draft algorithm for best wood cut search.
He will try implement a Convolutional Deep Belief Network based on the paper.

Ho Wing Teng

He has prepared foreground and background ratio feature for Character Recognition Training [Binarized image - Characters from real world image, Edge image - Characters from real world image, Binarized image - Artificial Characters, Edge image - Artificial Characters].
He has been testing on the spatial histogram (12th week).
He will be continuing on testing on the spatial histogram.

Richard Ng

He has read about the correlation filter based iris matching.
He is implementing the correlation filter of iris matching.
He will enhance on the iris recognition algorithm.

Chee Wei Lee

He has finish prepare a set of HOG feature with min_w =4 and min_h =4 with total of 27k features.
He has integrated the f_score calculation with the AdaBoost in selecting the threshold for f-score feature selection method. The selection threshold started with predefined possible threshold values and randomly spit the dataset into training and validation set. Then, calculate the test error on the validation set for these predefined threshold. The group of features with the lowest test error are use as the features for further training.
He will test the current and the new set of hog feature with this f_score selection and compare the result of with means square.

Ling Toh Koh

He has installed Dlink USB wireless driver into PICO820 and successful remote the PICO820 using NXclient tools with Yi Qian.
He has finish Master assignment 2 proposal development in chapter Research Methodologies and Development Tools which need to polish the English sentence.
He has link up the line can camera into PICO820 with 1 Gigabit Lan cable through the wireless remote.
He has been polishing English sentence on his assignment 2.
He will be continuing remote the PICO820 to control the microcontroller and line scan camera at parallel from time to time.

Yi Qian

He has successfully linked the line scan camera with MiniCpu (PICO 820).
He has installed Dlink USB wireless driver into Minicpu so to allow other pc to remote it.
He is configuring the File Access Permission and MTU on Ubuntu.
He will link the microcontroller to minicpu and complete the whole system within these 2 days.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Dr. Kai-Fu Lee Leaves Google, Starts Innovation Works

2009-09-07T18:15:00.001+08:00

Dr. Kai-Fu Lee Leaves Google, Starts Innovation Works
Supported by Industry Giants, Plans $115 Million Investment to Enable Chinese Entrepreneurs Develop Next Wave of High-Tech Companies in China

Beijing (September 7, 2009) – Innovation Works, a business creation platform focused on establishing the next wave of Chinese high-technology companies founded by Dr. Kai-Fu Lee, launches today. This announcement follows news of Dr. Lee’s resignation as Google Vice President and President of Google Greater China and introduces a new business paradigm into the Chinese start-up environment.
Led by Dr.Lee, Innovation Works will concentrate on Internet, mobile Internet, and cloud computing technology advancements targeted at the Greater China market, and build “dream teams” to collect, analyze, prioritize and execute on the most promising ideas. Innovation Works will accelerate an entrepreneur’s ability to prove ideas, obtain additional external funding, and then spin-off into an independent company.
Innovation Works will be funded with $115 million from an elite group of venture capital groups and investors including Steve Chen (Co-Founder of YouTube), Foxconn Technology Group, Legend Group and New Oriental Education & Technology Group. WI Harper Group is the lead venture capital investor.
“The Chinese entrepreneurial environment is still in its formative stage, with significant barriers for the early-stage entrepreneur: the lack of management experience and coaching, the reluctance of venture capitalists to invest in companies in the formation stage, and the lack of networking and experience to pull a company together,” said Dr. Kai-Fu Lee, Innovation Works founder, Chairman and Chief Executive Officer. “These barriers all contribute to a dearth of high-tech start-ups in China. Innovation Works is matching entrepreneurs, engineers, ideas, and capital with a unique business model that improves success rates and speeds time-to-market.”
Dr. Lee continued, “Innovation Works will be the de-facto institution for launching the most promising technology ideas in China. Through the rigorous development and testing of prototypes, and identification of a ‘founding executive’ to lead the venture, Innovation Works will provide capital, manpower, legal, financial and IT support. Our commitment is to mentoring and supporting the next-generation of Chinese entrepreneurs so that they can focus on building great products without distraction.”
Said Steve Chen, co-founder of YouTube, “As a co-founder of YouTube and PayPal, I deeply believe that passion and dedication are critical success factors of start-ups. Every time I visited China, I am encouraged by the enthusiasm and commitment shown by the entrepreneurs there. I am convinced that Innovation Works will provide the best platform for these entrepreneurs to realize their dreams.”
Terry Gou, Chairman of Foxconn, said: “I have known Dr. Lee since he worked for Apple, and I have followed Dr. Lee’s brilliant work from Apple to Microsoft to Google. I am confident that his creative ideas and charismatic leadership will be fully unleashed at Innovation Works. I am very pleased to be one of the founding investors of his new venture.”
Beijing Vice Mayor Gou Zhongwen endorsed Dr. Lee’s move, saying, “I hope Innovation Works will build on the advantages of Beijing Zhongguancun Science Park, China’s Silicon Valley, to foster more entrepreneurs, valued Chinese brands, intellectual property and high-end information technology. With Dr. Lee’s rich experience in technology innovation, Innovation Works will undoubted play a significant role in the next wave of advancements.”
Mr. Liu Chuanzhi, Chairman of Legend Group, said, “With the resilience of the Chinese economy, the abundance of late-stage funding, the newly established China Growth Market Second Board, the emerging 3G and cloud computing, and the exceptional talent pool in China, this is the perfect timing for Dr. Lee’s Innovation Works platform.”
Added WI Harper Chairman Peter Liu, “Every year Innovation Works will spin-off three to five companies and WI Harper is delighted to help those companies build an institutional investor syndicate to grow and flourish. Dr. Lee’s experience positions him perfectly to guide these new ventures to successful outcomes, and in doing so deliver value to all shareholders.”
More about Dr. Kai-Fu Lee:
Dr. Lee, 47, was born in Taiwan and grew up in the United States. He holds a B.A. from Columbia University (summa cum laude, 1983), and Ph.D. from Carnegie Mellon University (CMU). At Carnegie Mellon, he developed the first speaker-independent speech recognition system that was selected as the “Most Important Innovation” in 1988 by BusinessWeek. In 1989, he developed an Othello-playing computer program that defeated the world’s human champion.
After teaching for two years at CMU, Dr. Lee joined Apple in 1990, where he became the Vice President of its Interactive Media Group, which developed QuickTime and other technologies. In 1996, Dr. Lee joined Silicon Graphics Inc. (SGI) where he served as Vice President of the Web Products Division, which developed a $200 million product line of web servers, and also started Cosmo Software.
In 1998, he joined Microsoft where he built from the ground up Microsoft Research Asia, one of the world’s top research labs. MIT Technology Review calls this research lab “the hottest computer science laboratory.” Dr. Lee was promoted to Corporate Vice President in 2000, and he returned to the United States to co-lead Microsoft’s .Net effort.
In the fall of 2005, Dr. Lee joined Google to become a Vice President and President of Google Greater China.
During his tenure at Google, Dr. Lee built a 700-person organization with many successes. In the past three years, Google China’s market share nearly doubled from 16.1% to 31.0% (Analyses International). Google Map, Google Mobile Map, Mobile Search, and Google Translate have all become #1 in the China market this year.
In addition to his accomplishments in high technology, Dr. Lee has been devoted to fostering the professional growth of Chinese technology students. He has given hundreds of lectures to about 500,000 students, and has written seven public letters and three books for Chinese students. His first book, Be Your Personal Best, sold over one million copies and is still on many best-seller lists. Dr. Lee has built his own website for Chinese students (www.5xue.com) and answered thousands of questions from students. Dr. Lee is one of the most popular figures with young people in China.

About Innovation Works:
Founded in September 2009 and led by Dr. Kai-Fu Lee, Innovation Works is a business creation platform focused on establishing the next wave of Chinese high-technology companies and mentoring the next-generation of Chinese entrepreneurs. Innovation Works will concentrate on Internet, mobile Internet, and cloud computing technology advancements targeted at the Greater China market, and build “dream teams” to collect, analyze, prioritize and execute on the most promising ideas. Innovation Works matches entrepreneurs, engineers, ideas, and capital with a unique business model that improves success rates and speeds time-to-market.

Media Contact:
Samantha Chen
press@innovation-works.com
Tel: 13911631522

FICT Doctoral Presentation Session (Sept 8th, 1-4pm, PD206)

2009-09-05T21:33:00.001+08:00

Dear all,

We would like to invite you to a presentation session by our doctoral students who joined us this June. A total of 5 topic areas ranging from technical computer science materials to social information systems subjects will be presented. We plan for this session to be informal so everybody can participate in discussing, commenting, and providing suggestions to the presenters. This will be also a good "gathering and of course relaxing" session after 14-weeks of teaching. We know it is a long session, so we will prepare refreshments : ).

The schedule and the venue are as followed:

Date/Time: Tuesday, Sept 8th, 2009 / 1 - 4pm
Venue: PD206, UTAR Petaling Jaya

See you.

regards,
sfw.

Presentation/Demo from Omron

2009-09-05T21:31:00.001+08:00

Dear all,

You are invited to attend presentation/demo from Omron on 10th September 2009 (Thursday) at 2pm. Mr Looi, from Omron will give a presentation/demo on their lastest technology such as RFID, Interactive touch panel, floor displays and 3D animation that currently as an emerging technology in the market place. This presentation may benefits towards your research area.

The presentation/demo schedule as followed:

Date : 10th September 2009 (Thursday)
Time : 2pm
Venue : PD206 (UTAR PJ Campus, PD Block)

Thank you very much~

Best regards,
Yunli Lee
FICT, UTAR

Weekly Standing Meeting 2009.08.26

2009-08-29T21:54:00.006+08:00

Hao Wooi Lim

He has run NN on wing teng’s moment features with and without the junk class. With junk class, accuracy is around 30% while without junk class perform better – 65%. The train set has also been equalized to have same number of training instances for every class by random replication.
He has run pairwise binary classifier on wing teng’s moment features and obtain some results.
He has obtained a set of real-world images with rear view of vehicles from Caltech site.
He has obtained many free open source codes on CRF, HMM, LDA and DP.
He has successfully solved the light_speed matlab toolbox issues using MATLAB 2009a. The toolbox provides some common functions like repmat that is faster than the one provided in MATLAB.
He is currently working on his slides for the Research Methodology class. It is around 40% complete.
He is currently working on the proposal 2 for the Research Methodology class.
He will continue to work on a draft algorithm for best wood cut search.
He will try implement a Convolutional Deep Belief Network based on the paper.
He will continue working on the slides for the Research Methodology class.
He will continue working on proposal 2 for the Research Methodology class.

Ho Wing Teng

He has prepare another set of artificial characters [previously character training and testing set were obtained from the real world images’ vehicle’s license plate characters]. Each character is different in font type, shape, boldness, and rotation, brightness, for both training and testing set.
He has trained on the artificial characters [training set] and test on the artificial characters [testing set].
Prepare 4-corner foreground to background ratio feature on character recognition.
He will be continuing on testing on the spatial histogram.

Richard Ng

He has implemented the Wiener deblurring filter, entropy enhanced filter and denoise filter. It can be used effectively only when the frequency characteristics of the image and additive noise are known.
He is implementing the cross-correlation of the different rows of iris feature.
He will enhance on the iris recognition algorithm.

Lee Chee Wei

He has run the experiment on HOG features to compare with the result of NN and AdaBoost classifier. He has found that result of NN and AdaBoost with this feature set of HOG not perform well especially on the test set 2.
He has completed implement the f-score on the HOG feature selection.
He is currently experimenting on the feature selected from the f-score on the NN and AdaBoost.
He will create another set of HOG features (smaller in size) to test on the performance.

Ling Toh Koh

He is currently working on controlling the CSIBOT's server motor direction into few module.
He is preparing the power slides show for Research Methods subject.
He will be working on implement the module into PICO820.
He will be working on Assignment 2 which submit on next wednesday.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.08.19

2009-08-23T22:25:00.005+08:00

Tou Jing Yi

He has attended ICNC-FSKD 09 and presented 2 papers.

Lim Hao Wooi

He has coded a generic framework for testing character recognition and to generate Confusion matrix.
Ha has been reading up on clustering technique using Infinite Gaussian Mixture Model (IGMM) and has downloaded and played with Michael Mandel and Jacob Eisenstein’s IGMM implementation on matlab.
He has been reading up on a paper on object recognition using Transformed Dirichlet Process and downloaded its code and run some simple demo.
He has been reading on a paper on Convolutional Deep Belief Network.
He is currently working on the best angle of rotation searching part as first stage in wood cutting optimization.
He will continue to work on a draft algorithm for best wood cut search.
He will continue to try 4 corner foreground to background ratio feature on character recognition.
He will try implement a Convolutional Deep Belief Network based on the paper.
He will start preparing slides and proposal 2 for research methods.

Ho Wing Teng

He has run training on the character dataset (binarized Character Image) with FANN.
The testing on the training samples shows a low result, need to refine the training parameters.
He has been testing on the spatial histogram (10th week).
He will be continuing on testing on the spatial histogram. He will be testing each modules in the program.

Richard Ng

He is reading on the different types of image enhancement algorithm.
He is implementing the image enhancement algorithm.
He will enhance on the iris recognition algorithm.

Chee Wei Lee

He has tested the HOG feature in Adaboost (CART classifier, linearly combine single stage) on the number of selected feature with the number of iteration (number of weak classifier). Experiment ranging from the 110,165,220...to 880 feature from total of 5184 feature and tested in 500 iteration each for 3 difference. (With sign as the classificaiton)The result show that
(1) the test error (Avearge of error) is decreasing when the number of iteration increase.
(2) The slope of the test error drop faster when number of feature is higher.
(3) Gentle and Real Adaboost have better result compare to modest AdaBoost.
He has writing a program to test on the real sample. The performance of the detector on an negative image with only 18x36 scale window size sliding through whole image have 0.16 error rate(276/1726). On positive sample have 0.52 error rate (585/1237). He has come out some possible problems:
1) Make sure do not have coding error. (Check selected feature is match with is feature value. Verified, correct)
2) The selected feature are only base on one dimension of the feature..means need few dimension (weighted samples) or (boostrapping).
3) New AdaBoost classifier itself. (check with SVM)
4) Dataset itself (check with other dataset)
5) Feature selected it self.(check what is selected feature, feature valid or not, by using human eye, size of the feature, position )
6) Feature pool it self. (feature selected method, means distance, f-score, AdaBoost selection)
He is updating the pedestrian dataset in lab provided by a paper "A New Pedestrian Dataset for Supervised Learning".
He will examine each effect above and currently focus on the factor number 5.
He will modify the training classifier base on the threshold(confident of the weak classifier) and to the false positive and true positive rate.

Sim Chee Choong

He is doing report 1.
He is trying to broadcast voice message to another PC.

Vincent

He has been double checking the report 1.
He has been working on report 2.
He has been working on the presentation slide.
He will be continuing working on human motion detection.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

MIMOS Research Symposium : Recent Developnments in Human Motion Analysis

2009-08-21T17:14:00.001+08:00

Recent Developments in Human Motion Analysis

Chan Chee Seng, PhD, MIET, MIEEE

Centre for Multimodal Signal Processing

MIMOS Bhd.,

Technology Park Malaysia

ABSTRACT

Visual surveillance in dynamic scenes, especially for humans and vehicles, is currently one of the most active research topics in computer vision. It has a wide spectrum of promising applications, including access control in special areas, human identification at a distance, crowd flux statistics and congestion analysis, detection of anomalous behaviours, and interactive surveillance using multiple cameras, etc. In general, the processing framework of visual surveillance in dynamic scenes includes the following stages: modelling of environments, detection of motion, classification of moving objects, tracking, understanding and description of behaviours, human identification, and fusion of data from multiple cameras.

In this presentation, first of all I will review the recent developments and general strategies of all these stages. Secondly, I will show how some of my earlier research fits into this area. Finally, I will analyse possible research directions, e.g., occlusion handling, a combination of two and three-dimensional tracking, a combination of motion analysis and biometrics, anomaly detection and behaviour prediction, content-based retrieval of surveillance videos, behaviour understanding and natural language description, fusion of information from multiple sensors, and remote surveillance.

SPEAKER's PROFILE

Chee Seng Chan received his Ph.D. degree in Computer Science from the University of Portsmouth, Portsmouth, U.K., 2008. He previously held research appointment at the Institute of Industrial Research, University of Portsmouth, U.K. His current research interests include human motion analysis, medical image analysis, fuzzy qualitative reasoning, knowledge-based theory, and applications.

Date : 4 September 2009 (Friday)

Time : 10.00am - 11.00am

Venue : MIMOS Berhad, TPM, Kuala Lumpur.

While no registration is mandatory, it would be appreciated if you can kindly inform us of your participation on or before 2^nd September for our records and preparations.

Warmest regards.

Weekly Standing Meeting 2009.08.12

2009-08-15T15:19:00.005+08:00

Tou Jing Yi:

He has prepared the slides for ICNC 2009.
He will be attending ICNC-FSKD 2009.

Lim Hao Wooi:

Has has conducted some more experiments with SGQ-QN and BFGS learning algorithm.
He has tried character recognizer based on ratio of background to foreground pixels using SVM but did not obtained very good results.
He has finish working on a license plate detector based on his last paper but one without the need to use window scanning.
He is currently working on a character recognizer based on ratio of background to foreground pixels on 4 quadrants based on the character the centroid. The code to generate image from MNIST raw data set is done and the calculation of centroid based on image moments is also done.
He will format CVIS-009 (2nd week).
He will start work on an artificial real world images with license plate generator (2nd week).
He will start work on Wood Cutting optimization program.

Ho Wing Teng:

He has finish literature review on calculating moment feature for Character Recognition.
He has created Moment Feature Extractor for Character Recognition.
He has been testing on the spatial histogram (10th week).
He has been training on the Character Recognizer with Moment Feature
using FANN.
He will be continuing on the testing on the spatial histogram.

Richard Ng:

He has implemented the topological feature using euler code. The EER is 21.54%.
He has implemented using gray code feature. The EER is 21.46%.
He will enhance on the iris recognition algorithm.

Lee Chee Wei:

He has finish preparing the Feature Value on the all DC Dataset from the total 5184 features in matlab data by restructuring it as 500 features in a portion.
He had found the code for AdaBoost in matlab and has been using it to run some demo.
He had researched on some method of feature selection. He decide to use the F-score to select the feature and compare their performance with the mean distance. Another methods using for feature selection are GA, Mutual Information- such as Criteria of Max-Dependency, Max-Relevance and Min-Redundancy,PSO, Rough set theory or hybrid of the methods.
He had finish prepare the feature value from the total 110 selected feature based on the mean distance. Total of 10 features with highest mean distance value selected from 11 portion of the data value.
He will coding the selection of the feature by using F-score.
He will obtain the result of the HOG feature by using feature selected by mean distance and F-score. Difference experiment will be run on the diff number of the features and find the optimal setting.

Sim Chee Choong

He is trying to broadcast voice message to another PC.

Ling Toh Koh

He has submitted the SPICT paper with the title name “ Human detection using Motion Feature Abstract and Adaptive Boosting “ which are being check by Wing Teng.
He has been trying to get the power supply for Basler line camera and PICO 820 mini pc at Jalan pasar to ensure power adapter voltage and current doesn’t affect the functioning of both devices.
He will be continuing working on CSI-BOT with Kenny and Yi Qian.

Lab Issues:

Iris Sensor (Richard)
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Simon)
CVIS Dataset (Simon)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon)
Lab Backup (Simon, Jing Yi)

Weekly Standing Meeting 2009.08.05

2009-08-08T01:59:00.007+08:00

Tou Jing Yi

He has submitted a review paper on recent texture classification techniques.
He is currently preparing for a ICNC conference.

Lim Hao Wooi

He has submitted a joint journal on license plate detection using adaboost and unigram SIFT SVM.
He has submitted a paper also on license plate detection using a bigram SURF features.
He has played with 3 different online linear SVM learning algorithm (Stochastic Gradient Descent, Stochastic Gradient Descent with Quasi-Newton and BFGS) using the LIBSGDQN library. The findings shows that SGDQN is the fastest of all 3 and could sometimes converge in just 1 epoch.
He has added some features to the LIBSGDQN library.
He will be testing on character recognition problem using ratio of background to foreground pixels.
He will be working on a possible extension to his recent paper on bigram SURF features on license plate detection.
He will be working on a method to algorithmically generating realistically-looking license plate images.

Ho Wing Teng

He has helped reviewed Toh Koh's and Kelvin's paper.
He has been, and will continue to testing on spatial histogram (9th week).
He has been working on an extension to the "2-stage license plate detection" paper by incorporating character recognition.

Richard Ng

He has read another 3 paper on performance enhancement for degraded iris images.
He will propose a way to enhance his master thesis on iris recognition.

Lee Chee Wei

He has submitted his paper on pedestrian detection.
He has found the sample code for pyramid of HOG and has modified to match with method in the paper.
He is now coding the HOG feature for the HOG function in Matlab.
He will prepare the training data files to train in openCV Boosting function and SVM.
He will train the boosting and SVM classifier for this feature.

Sim Chee Choong

He has completed the configuration for Unreal Media Server and Unreal Live Server.
He is trying to broadcast voice message to another PC.
He is reading the Unreal Media Server tutorial and SDK.
He will continue testing on how to send voice message to other PC.

Lee Jun Yi

He is trying to complete his program for submission on this Friday.
He will prepare the documentation once his program is approved.

Kping Chin

He have installed visual studio and try out some basic coding for OpenCV.
He is reading the OpenCV reference.
He will be trying to use camera to capture image using OpenCV.

Lab Issues:

Iris Sensor (Richard)
Chair replacement
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Kenny)
CVIS Logo (Kenny, Kelvin)
CVIS Dataset (Jing Yi)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon, Jing Yi, Kenny)
Lab Backup (Jing Yi, Kenny)
NAS (Jing Yi, Richard, Kenny)

Weekly Standing Meeting 2009.07.29

2009-07-29T11:40:00.002+08:00

Jing Yi:

He has prepared the C Plugin for Covariance Matrix using MATLAB.
He has been working on SPICT and WAIC website and pamphlets (14th week).
He has been writing a review paper on texture classification (7th week).
He will be finalizing on the review paper on texture classification.
He will be obtaining and testing the labelling tool for WoodOPT.

Wing Teng:

He has been testing on the spatial histogram (8th week).
He has been running experiments on training multiple character detector.
He will be continuing on testing on the spatial histogram.

Chee Wei:

He has attended the Workshop on Computer Vision for video surveillance.
He has trained the feature value in the Adaboost function and run the testing on training feature.
He has been learning to use some tools in the MatLab practically.
He will be preparing the program that will save the feature value for training in dat file (binary file).
He will be restructuring the code on feature selection to calculate part by part to solve the memory leak problem.

Toh Koh:

He has attended the Hari pengguna 2009 Malaysia at Kuantan.
He has been spending time on CUDA tutorial and papers (4th week).
He has been editing the a conference paper with title “Human Detection Using Motion Feature and Adaptive Boosting” (2nd week).
He will be continuing on writing the paper for a conference.
He will be attending the INSEP CCNA & MCTIP course.

Simon:

He has prepared the GUI for the demo purposes for online and offline face detection.
He has been working on the conference paper on N-gram model on SURF for Object Recognition (due by this week).
He will be working on a joint paper on 2-stage License plate detection for Gabor covariance matrix (due by this week).
He will be working on the journal paper by adding more explanation for the SIFT portion and improving the classifier speed wise (due by this week).
He will be reporting on the status of the infected lab computers.

Junyi:

He has completed the analysis on the processing speed.
He has been working on the xml and database.
He will be completing the xml and database.
He will be attempting to further optimize the speed of the process.

Vincent:

He has been reading the OPENCV E-Book (4th week).
He has been working on the motion detection code.
He has been checking the problem with the OpenCV path.
He will be continuing on reading the OPENCV reference book.

Lab Issues:

Iris Sensor (Richard)
Chair replacement
Air ionizer (Kenny)
Water supply
Experiment Data Backup (Simon, Jing Yi)
Conference Publications (Simon, Jing Yi)
Printer Toner (Kenny)
CVIS Logo (Kenny, Kelvin)
CVIS Dataset (Jing Yi)
24/7 Pass (Chee Wei)
Google Sites (Simon, Jing Yi)
Virus Attack (Simon, Jing Yi, Kenny)
Lab Backup (Jing Yi, Kenny)
NAS (Jing Yi, Richard, Kenny)

Weekly Standing Meeting 2009.07.22

2009-07-22T12:12:00.001+08:00

Jing Yi:

He has been working on SPICT and WAIC website and pamphlets (13th week).
He has been writing a review paper on texture classification (6th week).
He will be continuing on writing the review paper.
He will be obtaining and testing the labelling tool for WoodOPT.
He will be preparing a C Plugin for Covariance Matrix using MATLAB.

Wing Teng:

He has written his thesis for Chapter 3.
He has been testing on the spatial histogram (7th week).
He will be continuing on testing on the spatial histogram.
He will be working with Simon on implementing Spatial Histogram with SVM.
He will be working on the extended journal.
He will be working on a conference paper on new features license plate detection.

Chee Wei:

He has managed to train the HOG feature selected by the feature selection method on the boosting algorithm in machine library in openCV.
He will be attending the Workshop on Advanced Techniques for Computer Vision in Video Surveillance.

Toh Koh:

He has completed enhancing his master proposal 1.
He has completed installing the CUDA driver, CUDA toolkits and CUDA SDK and has confirmed that cvis-010 NVIDIA graphic card supports CUDA.
He has been spending time on CUDA tutorial and papers (3rd week).
He has been editing the conference paper with title “Human Detection Using Motion Feature and Adaptive Boosting”.
He will be continuing on researching the CUDA and trying to implement the trained classifier in CVIS-010.
He will be continuing on writing the paper for a conference.

Simon:

He has completed the master proposal.
He has been reading 3 papers relevant to Master on object recognition.
He will be working on new features to extend the "Two-stage License Plate Detection using Gentle Adaboost and SIFT-SVM".

Junyi:

He has implemented the buttons within the GUI.
He has been working on the XML configuration file and doing the event trigger method for the total count for cars.
He has been analyzing the processing speed for the pulling method.
He will be completing the event trigger method and the analyzing of its processing speed.

Vincent:

He has been reading the OPENCV E-Book (3rd week).
He has been redoing the literature review.
He has been working on the motion detection coding.
He will be continuing on reading the OPENCV reference book.

Chee Cheong:

He has completed his proposal on literature review.
He has been learning on how to use the Unreal Media Server.
He will be continuing on exploring the Unreal Media Server.

Kok Ping:

He has completed his proposal on Literature review.
He has been studying OpenCV and using the software.
He will be starting to write the program.

Lab Issues:

Iris Sensor (Richard)
Chair replacement
Air ionizer (Kenny)
Water supply
Experiment Data Backup (Jing Yi)
Conference Publications (Jing Yi)
Printer Toner (Kenny)
CVIS Logo (Kenny, Kelvin)
CVIS Dataset (Jing Yi)
24/7 Pass (Chee Wei)
Google Sites (Jing Yi, Kenny)
Virus Attack (Jing Yi, Kenny)
Lab Backup (Jing Yi, Kenny)
NAS (Jing Yi, Richard, Kenny)

CCIS Seminar: How to use Grid technology in daily life and science

2009-07-20T07:58:00.001+08:00

Dear all,

You are cordially invited for a talk on grid computing by Dr Cheng Wai Khuen. Below are the details of the talk.

Title : How to use Grid technology in daily life and science
Date   : 27 July 2009 (Monday)
Time : 3.30 p.m.
Venue
   UTAR Perak Campus  : Conference Room, 1st Floor, Block A
   UTAR PJ Campus : Conference Room, 1st Floor, PD Block
Presenter   : Dr. Cheng Wai Khuen (Michael)
Abstract
Grid computing has emerged as a new paradigm for high performance computing. One of the objectives of Grid is to enable users to access remote resources and federate distributed computers for large-scale computational processes. Nowadays, the term of GRID is still a "Myth" for many non-computer science literate researchers. We see some degree of confusion in the IT industry since various terminologies such as on-demand computing, cloud computing, service oriented architecture and virtualization are bandied. This presentation will discuss on these issues and helps researchers to differentiate among the jargons. Some current Grid related projects which have been conducted globally in different domains are presented. It helps to inspire more researchers to apply this technology in the appropriate area for our daily life and science development. Some guides and concerns are presented to those interests in utilizing Grid for technology advancement.

Regards,
Mariam Nainan
Assistant Professor
Faculty of Information and Communication Technology
Universiti Tunku Abdul Rahman

CVT Seminar: Vehicle Safety and Injury Prevention by Prof Brian Fildes

2009-07-17T14:46:00.002+08:00

Date: 20th July 2009, Monday

Time: 3-4pm
Venue: DK1, UTAR KL Campus, Setapak

Title: Vehicle Safety and Injury Prevention

Speaker: Prof Brian Fildes

Biography: Professor Brian Fildes is Associate Director of MUARC Europe, based at the Monash University Centre in Prato, Italy. He has a Ph.D. in Psychology and qualifications in Engineering and his speciality is Human Factors research. He has particular interests in occupant protection, speeding, driver perception and injuries to older people, both on the road and in the home. Brian joined the Centre soon after it was established in 1987 and has developed and manages a team of researchers and graduate students working on a number of research projects. He has been instrumental in helping government agencies implement a number of new injury prevention countermeasures and programs as well as evaluating real world crash performance for the automotive industry.

All students and staff are invited

Collaboration Discussion: 2-3pm, 20th July 2009, 6th floor meeting room, all
staff are invited.

Thanks.

Regards,
Dr Tan Ching Seong 陈清祥

Weekly Standing Meeting 2009.07.15

2009-07-15T11:28:00.002+08:00

Jing Yi:

He has been working on SPICT and WAIC website and pamphlets (12th week).
He has been writing a review paper on texture classification (5th week).
He will be continuing on writing the review paper.
He will be updating the paper for a conference.
He will be obtaining and testing the labelling tool for WoodOPT.

Wing Teng:

He has been testing on the spatial histogram (6th week).
He has been working on Chapter 3 (3rd week).
He will be continuing on testing on the spatial histogram.
He will be continuing on Chapter 3.
He will be working with Simon on implementing Spatial Histogram with SVM.
He will be working on the extended journal.

Chee Wei:

He has analysed the boosting function in Machine Learning library of OpenCV and has run the demo on character recognition by using the training data provided by OpenCV.
He has been running the experiment which provides all the features to the system to construct the boosted tree classifier.
He will be doing the analysis and reading on the feature extraction methods.

Toh Koh:

He has completed his master proposal 1 and ready to submit.
He has been spending time on researching about the NVIDIA CUDA (2nd week).
He has been researching on collecting Haar features or other features value in txt form.
He will be continuing on researching the CUDA and attempting to use CVIS-010 PC.
He will be continuing on collecting features, and later scale and train them in SVM.

Simon:

He has written a new section on an extended journal.
He has provided a short guidance using LIBSVM.
He has been writing his Master Proposal 1’s Literature review (2nd week).
He has been training the Manifold Sculpting on MNIST (2nd week).
He will be starting to write the paper for a conference.

Vincent:

He has completed the project formulation.
He has been reading the OPENCV E-Book (3rd week).
He will be starting to write a program that detects the intruders.

Lab Issues:

Iris Sensor (Richard)
Chair replacement
Air ionizer (Kenny)
Water supply
Experiment Data Backup (Jing Yi)
Conference Publications (Jing Yi)
Printer Toner (Kenny)
CVIS Logo (Kenny, Kelvin)
CVIS Dataset (Jing Yi)
24/7 Pass (Chee Wei)
Google Sites (Jing Yi, Kenny)
Virus Attack (Jing Yi, Kenny)
Lab Backup (Jing Yi, Kenny)
NAS (Jing Yi, Richard, Kenny)

Weekly Standing Meeting 2009.07.08

2009-07-08T14:38:00.001+08:00

Jing Yi:

He has been working on SPICT and WAIC website and pamphlets (11th week).
He has been writing a review paper on texture classification (4th week).
He will be continuing on writing the review paper.
He will be updating the paper for a conference.

Wing Teng:

He has been testing on the spatial histogram (5th week).
He has been working on Chapter 3 (2nd week).
He will be continuing on testing on the spatial histogram.
He will be continuing on Chapter 3.
He will be working with Simon on implementing Spatial Histogram with SVM.

Chee Wei:

He has completing the paper draft.
He has corrected the error found in the feature value function.
He has studied the openCV Boosting function and went through the haartraining coding in openCV to find out the required information needed to integrate his own created feature value in vector format to the openCV for the training.
He has been reading a paper on the Histograms of Oriented Gradient for Human Detection.
He will be researching for other feature selection method for selecting the features.

Toh Koh:

He has completed the demonstration of the CSIBOT with Kenny and Yi Qian.
He has shared the knowledge of Adaboost training with the Yi Qian and Kelvin.
He has been spending time on reading a paper entitled “Detection using SVM and Multi-feature combination”.
He has been spending time on researching about the “CUDA” that was built by NVIDIA.
He will be continuing on researching the CUDA if it might be implemented for real time detection.
He will be continuing on reading others papers regards HOG.

Simon:

He has completed his Master Proposal 1 except for the Literature review.
He has modified the Waffles Machine learning library to run Manifold Sculpting on MNIST.
He has been writing his Master Proposal 1’s Literature review.
He has been training the Manifold Sculpting on MNIST.
He will be compiling results for the Manifold Sculpting training on MNIST.
He will be starting to write the paper for a conference.

Vincent:

He has completed the problem statement.
He has completed the literature review.
He has been working on the project formulation.
He has been reading the OPENCV E-Book (2nd week).
He will be starting to explore the development tool.
He will be continuing on reading the OPENCV reference book.

Lab Issues:

Iris Sensor (Richard)
Chair replacement
Air ionizer (Kenny)
Water supply
Experiment Data Backup (Jing Yi)
Conference Publications (Jing Yi)
Printer Toner (Kenny)
CVIS Logo (Kenny, Kelvin)
CVIS Dataset (Jing Yi)
24/7 Pass (Chee Wei)
Prepare Pedestrian Dataset (Chee Wei)
Google Sites (Jing Yi, Kenny)
Virus Attack (Jing Yi, Kenny)
Lab Backup (Jing Yi, Kenny)
NAS (Jing Yi, Richard, Kenny)