高效视频编码帧内快速深度决策算法

Fast intra-depth decision algorithm for high efficiency video coding

针对新一代高效视频编码(HEVC)帧内预测中编码单元(CU)的编码深度选择过程中计算复杂度较高的问题,提出了一种基于空域相关性的帧内快速深度决策算法。首先,利用相邻已编码树单元(CTU)的深度通过线性加权得到当前CTU深度估计值;然后,对当前CTU深度估计值设置较为合适的深度双阈值提前终止编码树单元的划分或跳过CTU的某些深度,来缩小当前CTU的深度范围,从而减少不必要的深度计算。实验结果表明:与HM12.0相比,所提算法对比较简单的视频序列编码时间的减少比较明显,在亮度峰值信噪比(Y-PSNR)几乎不变的情况下(平均降低0.02 dB),编码时间平均减少了34.6%。此外,所提算法容易与其他算法进行融合,能进一步降低HEVC的帧内计算复杂度,最终达到实时传送高清视频的目的。

To reduce the high computational complexity of intra coding in High Efficiency Video Coding （HEVC）, a fast intra depth decision algorithm for Coding Unit （CU） based on spatial correlation property of images was proposed. First, the depth of the current Coded Tree Unit （CTU） was estimated by linearly weighting the adjacent CTU depths. Then appropriate double thresholds were set to terminate the CTU splitting process or skip some depths of CTU, thereby reducing unnecessary depth calculation. Experimental results show that compared with HM12.0, the proposed intra-depth decision optimization algorithm can significantly decrease the coding time of simple video sequence with only negligible drop in quality, when the Y-PSNR dropped by an average of 0.02 dB, the encoding time is reduced by an average of 34.6%. Besides, the proposed algorithm is easy to be fused with other methods to further reduce the computational complexity for HEVC intra coding, and ultimately achieves the purpose of real-time transmission of high-definition video.