为进一步提高分布式视频编码(distributed video coding,DVC)的压缩性能,针对离散小波变换域DVC,提出了基于分层细化的Wyner-Ziv解码算法。算法充分利用小波多尺度和多分辨率的特性,将边信息优化算法和高阶统计模型进行了深度融合。在比特层面上,通过边信息优化算法提升每一分解层高频子带的边信息质量,从而提高高阶统计模型中与边信息相关的两大特征的准确性,增强高阶统计模型在信源相关性挖掘和有效利用方面的性能,实现DVC压缩性能的提升。测试结果表明,与参考文献相比,基于本文算法的DVC系统压缩性能有明显提高。
In the Wyner-Ziv(WZ) video coding paradigm, a virtual correlation channel is assumed between the quantized source and the side information(SI) at the decoder, and channel coding is applied to achieve compression. In this paper, errors caused by the virtual correlation channel are addressed and an error concealment approach is proposed for pixel-based WZ video coding. In the approach, errors after decoding are classified into two types. Type 1 errors are caused by residual bit errors after channel decoding, while type 2 errors are due to low quality of SI in part of a frame which causes SI not lying within the quantization bin of a decoded quantized pixel value. Two separate strategies are respectively designed to detect and conceal the two types of errors. Simulations are carried out and results are presented to demonstrate the effectiveness of the proposed approach.
Yuan ChengzongHan ChuanzhaoWang YanZhang NingZhang ZhenZhu XinzhongLi XianQu Di
Side information has a significant influence on the rate-distortion（RD） performance of distributed video coding（DVC）. In the conventional motion compensated frame interpolation scheme, all blocks adopt the same side-information generation method regardless of the motion intensity inequality at different regions. In this paper, an improved method is proposed. The image blocks are classified into two modes, fast motion and slow motion, by simply computing the discrete cosine transformation（DCT） coefficients at the encoder. On the decoder, it chooses the direct interpolation and refined motion compensated interpolation correspondingly to generate side information. Experimental results show that the proposed method, without increasing the encoder complexity, can increase the average peak signal-to-noise ratio（PSNR） by up to 1~ 2 dB compared with the existing algorithm. Meanwhile, the proposed algorithm significantly improves the subjective quality of the side information.
