基于SHVC空间可伸缩的帧内快速编码算法

Fast intra algorithm for spatial scalability in SHVC

根据SHVC(Scalable High-efficiency Video Coding)空间可伸缩视频编码的特点,提出一种空间可伸缩视频编码的帧内快速算法。首先将视频分类,对不同视频序列采用不同的权值结合相关编码单元的深度进行预测,排除可能性较小的深度;然后利用残差系数的分布研究部分零块的提前终止判断条件;最后根据相关编码单元的方向模式预测可能的方向模式,排除可能性较小的方向模式。实验证明本文算法显著地提高编码速度,且编码效率损失较小。

According to the characteristics of spatial scalable video coding,this paper presents a fast intra algorithm based on spatial scalable video coding. First,videos are classified. Different weights are used for different video sequences,and the depth of the relevant coding unit is combined to make the prediction,to eliminate the depth of small possibility. Then the condition of early termination of partial zero-value block is studied by using the distribution of residual coefficient. Finally,according to the direction modes of the relevant coding unit,the possible direction modes are predicted,and the direction modes with less possibility are excluded.Experimental results show that the algorithm can significantly improve the coding speed and the loss of coding efficiency is very small.