The switching from analog to digital broadcast television is starting to relegate big picture
tube TVs with a mere 720x480 resolution towards the footnote. The general public is now
more aware of the vastly improved 1920x1080 resolution and exceptionally crisp images
that HDTV technology could offer. This success of HDTV will eventually spill over into the
video surveillance market and we foresee that it would not take long for the surveillance
industry to start demanding HDTV quality video footage for recording and remote
monitoring. On the other hand, the widespread adoption of H.264, for instance in digital
broadcast and Blu-Ray disc, also arouse the interests of adopting H.264 standard for use in
video surveillance, due to its superior compression efficiency without compromising video
quality. As a result, it is expected that HDTV and H.264 technologies are two main driving
forces for the growth in the video surveillance market. As a matter of fact, more and more
manufacturers like Axis, Sony, Samsung, etc. have already started to include megapixels
camera with H.264 capabilities into their products portfolio, indicating the drifts in market
expectations towards HDTV and the H.264 standard.
However, there is one major gap that holds back the propagation of HDTV and H.264,
which are now matured technologies, into the surveillance market. Essentially, the lacking
of efficient H.264 decoding of multiple video streams on an off-the-shelf workstation is the
main obstacle for HDTV and H.264 to move forward in the surveillance industry.
Although there are quite a few efficient software or libraries available for H.264 decoding,
but they are mostly designed and optimized for single video playback decoding, where
end-to-end video latency is never a concern. Video surveillance, in contrast, does require
efficient handling of multiple video live streams decoding with very low latency requirement.
As a consequence, the video surveillance industry is in need of a highly efficient decoding
technology that fuses this gap, so that the industry as well as CCTV users can enjoy the
benefits that HDTV and H.264 could offer.
In view of the aforementioned background and issues, we propose to address the video
decoding problem by offloading the video decoding as well as the post-processing tasks to
the General Purpose Graphics Processing Unit (GPGPU). Specifically, our proposed
solution aims at identifying only those decoding portions that fits best into the stream
processor architecture, and offload them from CPU to GPGPU. In this way, the GPGPU
serves as a video coprocessor and speed up can always be guaranteed. Besides, using
GPGPU as a coprocessor allows us to have a smaller granularity and larger freedom in
scheduling subtasks on GPGPU, so that multiple video streams can be scheduled to
decode efficiently with low latency, which is a basic and essential requirement in video
surveillance. The conceptual diagram is illustrated in Figure 1, which demonstrates that
execution time can be shortened if some portions of execution can be offloaded onto the
GPGPU.
Figure 1: Execution time reduction by offloading subtasks to GPGPU
[4] Feifeng Zheng, Francis Y.L. Chin, Stanley P. Y.
Fung, Chung Keung Poon and Yinfeng Xu,
“A Tight Lower Bound for Job Scheduling with
Cancellation”, Information Processing Letters,
Elsevier, 97(1):1-3 (January 2006)
[5] Wun-Tat Chan, Francis Y.L. Chin, Yong Zhang,
Hong Zhu, Hong Shen and Prudence W.H.
Wong, “A Dynamic Programming Approach Of Finding An
Optimal Broadcast Schedule InMinimizing Total Flow
Time”, Journal of Combinatorial Optimization,
Springer Netherlands,
11(2): 177-187 (March 2006) – a preliminary version
also appeared in the Proceedings of
COCOON 2005
[6] Francis Y.L. Chin, Marek Chrobak, Stanley P.Y.
Fung, Wojciech Jawor, Jiří Sgall and Tomáš
Tichý, “Online Competitive Algorithms for Maximizing
Weighted Throughput of Unit Jobs”,
Journal of Discrete Algorithms, Elsevier B.V.,
4(2):255-276 (June 2006) – a preliminary
version also appeared in the Proceedings of STACS
2004
[7] Ronald H.Y. Chung, Kenneth Kwan-Yee Wong,
Francis Y.L. Chin, K.P. Chow and S.C. Yuk,
“Generalized Motion and Edge Adaptive Interpolation
De-interlacing Algorithm”, WSEAS
Transactions on Computers 5(11) : 2544-2551
(November 2006) – a preliminary version
also appeared in the Proceedings of the 10th WSEAS
International Conference on
Computers 2006
[8] Chen Zhuo, Francis Y.L. Chin and R.H.Y. Chung,
“Automated Hierarchical Image
Segmentation Based on Merging of Quadrilaterals”,
WSEAS Transactions on Signal
Processing, Crete, Greece, Issue 8, Volume 2,
1063-1068 (August 18-20, 2006) - a
preliminary version appeared in the Proceedings of
the 6th WSEAS Int. Conf. on Signal
Processing, Computational Geometry & Artificial
Vision 2006 (CD, ISBN:960-8457-51-3 http://www.wseas.org)
[9] Keqiu Li, Hong Shen, Francis Chin, Weishi Zhang,
“Multimedia Object Placement for
Transparent Data Replication,” IEEE Transactions on
Parallel and Distributed Systems,
18(2):212-224 (February 2007)
[10] Wun-Tat Chan, Francis Y. L. Chin, Deshi Ye,
Guochuan Zhang, Yong Zhang, “On-Line
Scheduling of Parallel Jobs on Two Machines”,
Journal of Discrete Algorithms, 6(1):3-10
(March 2008) ISSN:1570-8667 – a preliminary version
appeared in the Proceedings of
AWOCA 2005
[11] Stanley P. Fung, Feifeng Zheng, Wun-Tat Chan,
Francis Y.L. Chin, Chung-Keung Poon and
Prudence W.H. Wong, “Improved on-line broadcast
scheduling with deadlines”, Journal of
Scheduling, 11 (4): 299 - 308 (August 2008)
ISSN:1094-6136, Kluwer Academic Publishers
Hingham, MA, USA - a preliminary version appeared in
the Proceedings of COCOON'06
[12] Keqiu Li, Takashi Nanya, Hong Shen, Francis Y.
L. Chin and Weishi Zhang, “An Efficient
Cache Replacement Algorithm for Multimedia Object
Caching”, International Journal of
Computer Systems Science and Engineering (to appear)
[13] Kwan-Yee K. Wong, Francis Y.L. Chin, Ronald H.Y.
Chung, K.P. Chow and S.C. Yuk,
"Motion and Edge Adaptive Interpolation
De-interlacing Algorithm", Proceedings of the 10th
WSEAS International Conference on Computers, Athens,
Greece, 1030-1035. (July 13-15,
2006)
[14] Feifeng Zheng, Stanley P.Y. Fung, Wun-Tat Chan,
Francis Y.L. Chin, Chung Keung Poon,
and Prudence W.H. Wong, “Improved on-line broadcast
scheduling with deadlines”,
Proceedings of the Twelfth Annual International
Computing and Combinatorics Conference
(COCOON 2006), Taipei, Taiwan, 320-329 (August
15-18, 2006)
[15] Chen Zhuo, Francis Y.L. Chin and R.H.Y. Chung,
“Automated Hierarchical Image
Segmentation Based on Merging of Quadrilaterals”,
The Processings of The 6th WSEAS
International Conference on Signal Processing,
Computational Geometry & Artifical Vision
(ISCGAV’06), Elounda, Crete, Greece, 135-140 (August
18-20, 2006) (CD,
ISBN:960-8457-51-3 http://www.wseas.org)
[16] Jacky S. C. Yuk, Kwan-Yee K. Wong, Ronald H. Y.
Chung, F. Y. L. Chin and K. P. Chow,
"Real-time Multiple Head Shape Detection and
Tracking System with Decentralized
Trackers", Proceedings of the 6th International
Conference on Intelligent Systems Design
and Applications (ISDA'06), Shandong, China, Vol.
II, 384-389 (October 16-18, 2006)
[17] X. He, N.H.C. Yung, K.P. Chow, F.Y.L. Chin,
R.H.Y. Chung, K.-Y. K. Wong, K.S.H. Tsang,
“Watershed Segmentation with Boundary Curvature
Ratio Based Merging Criterion,”
Proceedings of the 9th IASTED International
Conference on Signal and Image Processing
(SIP 2007), Honolulu, Hawaii, USA, 7-12 (August
20-22, 2007)
[18] Jacky S-C. Yuk, Kwan-Yee K. Wong, Ronald H-Y.
Chung, K. P. Chow, Francis Y-L. Chin,
Kenneth S-H. Tsang, "Object-based Surveillance Video
Retrieval System With Real-Time
Indexing Methodology", The International Conference
on Image Analysis and Recognition
(ICIAR 2007), Montreal, Canada, 626-637 (August
22-24, 2007)
