Xianglin Wang

FGS Coding with Adaptive Reference for Low-Delay Applications

This paper presents the design of a fine granularity SNR scalable video coder for low-delay applications. Specifically, temporal prediction is introduced into the FGS layer of closed-loop P-frames. The prediction is formed adaptively from the enhancement layer reference and the base layer reference, utilizing information coded in the base layer. This novel FGS coder can control drift that may occur due to partial decoding of the FGS layer, while at the same time achieving high coding efficiency. Effort is also spent on reducing the complexity of the new solution by using low-complexity motion compensation and by minimizing additional transform operations. With minimal increase in the complexity, the FGS coding performance of closed-loop P-frames can be improved by as much as 4.5 dB for one FGS layer on a base layer coded at QP 42.

Improved video coding with residual prediction for extended spatial scalability

Annex G has been added to the H.264/AVC video coding standard to improve the scalability of coded video. In this annex, residual prediction is one of the main coding tools that provide efficient inter-layer prediction with moderate coding complexity. However, when residual prediction is used in an extended spatial scalability scenario, it is reported that certain visual artifacts can appear, which results in a perceived quality degradation. This paper provides a detailed analysis of the problem. Based on the analysis, a solution that involves a modified mode decision process is suggested. With the proposed method, the reported visual artifact may be completely prevented. Meanwhile, the proposed solution does not require any change to decoder and therefore does not incur any overhead to decoder complexity.