Error-resilient surveillance video transmission based on a practical joint source-channel distortion computing model

• Published in: The Journal of supercomputing Vol. 73; no. 3; pp. 1017 - 1043
• Main Authors: Jung, Tae-jun, Jeong, Yo-Won, Lee, Hong-rae, Seo, Kwang-deok
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2017; Springer Nature B.V
• Subjects: Bandwidths; Coding; Compilers; Complexity; Computation; Computer Science; Computer simulation; Decision analysis; Distortion; Error analysis; Error correction; Interpreters; IP (Internet Protocol); Processor Architectures; Programming Languages; Public safety; Surveillance; Systems design; Test procedures; Training; Vibration; Video transmission; Forward error correction; Error-resilient surveillance video transmission; Joint source-channel coding; FEC code rate decision
• ISSN: 0920-8542; 1573-0484
• DOI: 10.1007/s11227-016-1762-3

In recent years, IP (Internet Protocol)-based video surveillance has widely been useful for post-event analysis and assisting the work of privacy protection and public safety. To support high-quality IP video surveillance, error-resilience techniques are important for surveillance system design, because video has more stringent requirements than general video transmission for packet loss, latency, and jitter. The optimal FEC (forward error correction) code rate decision is a crucial procedure to determine the optimal source and channel coding rates to minimize the overall picture distortion when transporting video packets over packet loss channels. The conventional FEC code rate decision schemes using an analytical source-coding distortion model and a channel-induced distortion model are usually complex and typically employ the process of model parameter training, which involves potentially high computational complexity and implementation cost. To avoid the complex modeling procedure, we propose a simple but accurate joint source-channel distortion model to estimate the channel-loss threshold set for optimal FEC code rate decision. Since the proposed model is expressed as a simple closed form and has a small number of scene-dependent model parameters, a video sender of the surveillance system using the model can be easily implemented. For training the scene-dependent model parameters in real time, we propose a practical test-run procedure. This method accelerates the test-run while maintaining its accuracy for training the scene-dependent model parameters. Using the proposed simple model and practical test-run method, the video sender can find the optimal code rate for on-the-fly joint source-channel coding whenever there is a change in the packet-loss condition in the channel. Simulations show that the proposed method can accurately estimate the channel loss threshold set, resulting in an optimal FEC code rate with low computational complexity.