空间可扩展视频编码中增强层上的帧内预测算法研究

Spatially Scalable Video Coding with Intra Prediction in Enhancement Layer

针对空间可扩展视频编码,提出了一种新的增强层上的帧内预测算法。该算法充分考虑了相邻层间所固有的空域相关性,在无法正确获取本层邻块预测值的情况下,利用基本层上插值放大后的解码,通过重构宏块来对增强层上的编码宏块进行帧内预测。此外,还在预测过程中,通过引入与图像纹理特征相关联的自适应加权因子对(Wv,Wh)来进一步提高新算法的编码性能。实验结果表明,与H.264标准中的帧内预测算法相比,新算法在保持原有码率和不增加额外的计算复杂度的前提下,可一定程度地提高亮度分量的PSNR值,因此可以作为空间可扩展视频编码中增强层上的帧内编码方案。

In this paper, a new intra prediction algorithm for enhancement layer in spatially scalable video coding is proposed. It takes the spatial correlation between adjacent layers into consideration, and makes use of the magnified decoded and reconstructed base layer block for intra prediction in the lack of neighboring predicting samples. Furthermore, two adaptive weight factors （ Wa, Wb ） related to the texture characteristic within an image are introduced to improve the coding efficiency of proposed algorithm. Experimental results show that the PSNR value of luminance component has been increased and both coding bit-rate and computation complexity are preserved. As a result, it can be used as an efficient intra coding scheme for enhancement layer in spatially scalable video coding.