AVS-M is part seven of AVS standard, which is the second generation source coding standard of China. Because the system resource is limited in mobile platform compared with AVS P2, AVS-M has lower complexity, and is suitable for mobile-multimedia applications. The performance of AVS-M is close to H.264/AVC in the case of low rate bitstream. It is very important to increase the real-time encoding and decoding speed in mobile platform. One way is to optimize the AVS-M source code; on the other hand, simplifying or masking some modules in AVS-M to increase the speed greatly, and simultaneously, the score decrease of perceptual evaluation is small. In this paper, considering the characteristics of mobile platform (such as the case of low resolution, limited system resource etc.), in order to compare the differences between objective and subjective evaluation, we proposed a testing scheme which covers most of the main modules in AVS-M. According to the testing results, we analyze the influence of different modules on objective and subjective evaluation.
[1]
Paul M. Chau,et al.
Objective Human Visual System Based Video Quality Assessment Metric for Low Bit-Rate Video Communication Systems
,
2006,
2006 IEEE Workshop on Multimedia Signal Processing.
[2]
Feng Wu,et al.
Overview of AVS video standard
,
2004,
2004 IEEE International Conference on Multimedia and Expo (ICME) (IEEE Cat. No.04TH8763).
[3]
Wenguang Jin,et al.
Optimization and Implementation of AVS-M Decoder on ARM
,
2007,
Fourth International Conference on Image and Graphics (ICIG 2007).
[4]
Peng Tang,et al.
Salient Moving Object Detection Using Stochastic Approach Filtering
,
2007,
Fourth International Conference on Image and Graphics (ICIG 2007).
[5]
Zhu Guang-xi.
Overview of AVS-M Video Coding Standard
,
2005
.
[6]
Yin Hanchun.
Performance Comparison between AVS and H.264 Video Codec Standards
,
2007
.