基于像素分组的帧内预测快速算法

为了降低HEVC帧内编码的复杂度,基于HM 12.0提出了一种帧内预测快速算法.首先将35种预测方向分成5组,然后基于像素提取PU块的梯度方向信息并匹配到其中一组,最后在RDO运算时根据统计信息适当降低了候选模式的数量.实验结果表明,与JCTVC最新源码HM 12.0相比,该算法平均能够节省20％的编码时间,峰值信噪比和码率变化不大.

Fast algorithm for intra prediction based on pixel grouping

A fast algorithm based on HM-12.0 is proposed in this paper to reduce the complexity of intra prediction. First, the direction of 35 kinds of prediction modes is divided into five groups, and the current PU is matched to one group based its gradient direction information which is extracted by the proposed algorithm. At last, the number of candidate modes for RDO is also reduced accordingly. The experiment on the last HM-12. 0 shows that,the proposed algorithm can save an average of 20% time for encoding,and the PSNR and the rate changed little.