A Hybrid Importance Metric for Fine-granularity Packet Classification of Delay-constraint Video Stream
(0) For a delay-constraint video stream over DifiServ networks, different video packets have unequal contributions to the overall quality. According to a certain sender-side delay, the video stream is firstly divided into a series of delay-constraint packet sets. By analyzing the available coding information in encoded frames and error-propagation effects in future frames, we proposed a hybrid importance metric to online estimate the GOP-level transmission distortion of each video packet in a delay-constraint packet set. Under different delay constraints, the hybrid importance metric can evaluate the relative importance of different video packets, and then assign a unique importance class to each packet in the delay-constraint packet set. The proposed metric can be aware of non-stationary video content and achieve the packet-by-packet classification accuracy. Simulation results have demonstrated the effectiveness of the proposed metric for fine-granularity packet classification of delayconstraint video stream.
Hao Liu Shu-guang Zhao Jian-min Qian
Engineering Research Center of Digitized Textile & Fashion Technology, Ministry of Education, Donghua University, Shanghai 201620, China
国际会议
2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)
上海
英文
10-13
2010-12-25（万方平台首次上网日期，不代表论文的发表时间）