Changsong Shen
 |
PhD
Candidate
Advisor: Dr. Sidney F. Fels and Dr. James J. Little
Human Communication Technologies Laboratory
Office: X509 ICICS building
Phone:
(604)822-4583
E-mail: csshen AT ece^ubc^ca
Department of Electrical and Computer Engineering
The
2332
Main Mall, Vancouver BC V6T 1Z4 Canada
|
Hi, welcome to my home page. Here is a brief introduction of myself. I got M.S. degree from Department of Computer Science at University of Science & Technology of
My research interests include Computer Vision, Real-Time System and Software Architecture. My current work focuses on designing Open Vision Library, which is a general software architecture model of computer vision facilitating for hardware acceleration, reusability and scalability.
Research
Open Vision Library (OpenVL)
Open Source Vision
Library (OpenVL) is a software architecture model designed to address hardware
acceleration, reusability, and scalability for video-based surveillance systems and other computer vision
applications. Our approach is
inspired by the success of OpenGL in promoting the development of hardware
acceleration for computer graphics. Our goal is to create OpenVL
as a standard interface for computer vision applications that can work
seamlessly on different software and hardware platforms supporting distributed
camera arrays. It allows users to easily recover useful information about multiple dynamic scenes quickly, and in a portable
manner across various software and hardware platforms.
Quality-of-View (QOV) Measurement
In this work, we propose a multi-camera video surveillance system with automatic camera selection. A new confidence measure, Quality-Of-View (QOV), is defined to automatically evaluate the camera’s view performance for each time instant. This measure takes into account view angle and distance from subjects. By comparing each camera’s QOVs, the system can select the most appropriate cameras to perform specific tasks. We also present an approach to determine the minimum number of cameras and their layout in a convex polygonal room under specific QOV constraints. Finally, we implement an experimental surveillance system, to confirm the stability of our algorithm and validate the critical underlying concepts of QOV.
Local Positioning System (LPS)
This research addresses
the architecture of a remote low-cost position sensing infrastructure for
ubiquitous computing. To reduce the cost, each tracked object carries an
inexpensive active tag that emits a modulated near-infrared signal, which is
received and decoded by a camera module. Each camera module consists of a CMOS
camera connected to a Field Programmable Gate Array
(FPGA) and Ethernet communication hardware. Remote sensing can be achieved by
means of Ethernet. The use of inexpensive FPGA to implement the most time- and
data-critical tracking algorithms provides a cost-effective platform with low
latency. The preliminary results suggest it is successful at tracking objects
and feasible for achieving medium fidelity tracking performance in both space
and time.
Recent Publications
Book Chapter
1.
Changsong Shen, James Little and Sidney Fels, "Towards OpenVL: Improving Real-Time Performance of Computer Vision Applications", Embedded Computer Vision, Springer 2008 (in press)
Conference Papers
1.
Changsong Shen, James Little and Sidney Fels, "OpenVL: Towards A Novel Software Architecture for Computer Vision", The Third IEEE Workshop on Embedded Computer Vision, In conjunction with CVPR'07, Minneapolis, Minnesota, Jun 19-23 2007
2.
Changsong Shen,
Chris Zhang and Sidney Fels, "A Multi-Camera Surveillance System that estimates Quality-of-View Measurement
", The IEEE International Conference on Image Processing (ICIP 2007), Accepted, San Antonio, Texas, Sep 2007
3.
Changsong Shen,
Steve Oldridge and Sidney Fels, "Open Source Vision Library (OpenVL) Based Local Positioning System", Proceedings of IEEE Conference Series on Advanced Video and Signal based Surveillance (AVSS 2006), pp.105-110, Sydney, Australia, Nov 2006
4.
Changsong Shen,
Baosheng Wang, Florian
Vogt, Steve Oldridge and Sidney Fels,
"RemoteEyes: A Remote Low-Cost Position Sensing Infrastructure for
Ubiquitous Computing", Proceedings of 1st International Workshop on
Networked Sensing Systems (INSS 2004), pp. 31-35,
Journal
Papers
1.
Changsong Shen,
Baosheng Wang, Florian
Vogt, Steve Oldridge and Sidney Fels, “RemoteEyes: A Remote Low-Cost Position Sensing Infrastructure for Ubiquitous
Computing”, Transactions of
the Society of Instrument and Control Engineers, Vol
E-S-1, Pages 85-90, 2005
2.
Changsong Shen,
Xiaopei Luo, “The
Technology of
3.
Changsong Shen,
Xiaopei Luo, Jianfang Zhu, “A Performance Test Method for
Real-Time Operating System”, Microcomputer Information, Apr 2001
4.
Changsong Shen,
Xiaopei Luo, Fang Xie, “Display Technology with Superposition of
Scrolling Screen and Static Image”, Microcomputer Development, Oct 2000
Poster
Presentations
1.
Changsong Shen,
Steve Oldridge, Amir Afrah, Gregor Miller, and Sidney Fels,
“Design and Implementation of an What-Oriented Open Vision Library”, IS 2008 : 18th Annual Canadian Conference on Intelligent Systems,
2.
Changsong Shen,
Steve Oldridge, Amir Afrah, and Sidney Fels,
“Open Source Vision Library (OpenVL) ”, IS 2006 : 16th Annual Canadian Conference on
Intelligent Systems,
3.
Changsong Shen
and Sidney Fels, “OpenVL Based Human Tracking System”, IS 2005 : 15th Annual Canadian Conference
on Intelligent Systems,
4.
Changsong Shen,
Baosheng Wang, Florian
Vogt, Steve Oldridge, and Sidney Fels, “RemoteEyes: A Remote Low-Cost Position Sensing Infrastructure for
Ubiquitous Computing”, IS 2004 : 14th Annual Canadian Conference on Intelligent Systems,
Personal
When I'm away from research, I like hiking, tennis or swimming. Now I have a new hobby: Yoga. J
|
|
|
Last modified by Changsong, May 20th, 2008 |