Abstract: This paper proposes an adaptive quantization algorithm named AQSC in transform-domain Distributed Video Coding(DVC). Only the important Discrete Cosine Transform(DCT) bands determined by the statistical characteristics of DCT coefficients are quantized and coded, while the less important ones are not quantized and coded, but replaced directly by the side information in the decoder side. Experimental results show that AQSC achieves rate-distortion improvements and reduces the decoding rates, compared to IST-TDWZ scheme.

Key words: transform-domain Wyner-Ziv video coding, Distributed Video Coding(DVC), Discrete Cosine Transform(DCT), statistical characteristics of DCT coefficients, sub-band grouping, adaptive quantization

摘要： 针对变换域分布式视频编码问题，提出一种自适应量化算法AQSC。根据视频帧离散余弦变换系数的分布特点，确定各子带的重要性，只对重要的子带进行量化及编解码，不重要的子带不进行量化及编码，且在译码时直接用边信息代替。实验结果表明，AQSC算法可提高分布式视频编码系统的率失真性能，减少译码码率，解码后视频序列的整体质量有所提高。

关键词: 变换域Wyner-Ziv视频编码, 分布式视频编码, 离散余弦变换, DCT系数统计特性, 子带分组, 自适应量化

CLC Number: 

• TP301.6