一种HEVC帧内快速编码算法

A fast intra coding algorithm for HEVC

摘要:

高效视频编码(HEVC)采用编码单元(CU)四叉树的 分割结构,相比H.264/AVC显著地提升了编码效 率,但却使编码复杂度急剧增加。为此,本文提出一种帧内快速编码算法。首先,根据视 频图像纹理复 杂度,提前判断是否进行最大编码单元(LCU)分割。然后,根据空域相邻CU的深度预测当前C U的深度范围, 跳过不必要的计算；最后,根据预测模式被选为最优预测模式的统计特性,去掉可能性小的 帧内预测模式。本文算法在HM14.0的基础上实现。 仿真结果表明,本文算法在全I帧模式下与HM14.0相比,帧内编码时 间平均减少38%,码率(BR)只增加1.41%,峰值信噪比(PSNR)只降低0.29dB,在保证编码性能和视频质量几乎不变的 情况下,本文算法降低了编码的计算复杂度。

Abstract:

High efficiency video coding (HEVC) adopts the coding unit with a quad-tree p artitioning structure that can improve the coding efficiency significantly,but this causes a dramatic incr ease in computational complexity compared with H.264/AVC.Therefore,this paper proposes a fast algorithm for intra coding.Firstly,the texture complexity of a largest coding unit (LCU) is used in advance to determine whethe r it should be split or not.Then,the depth range of the current coding unit is predicted based on depths of the spat ially nearby CUs and helps to skip the unnecessary calculation.Finally,based on statistical characteristics that the intra prediction modes are selected to be the optimal prediction modes,the intra prediction modes with small possibi lity are excluded.The experimental results show that the proposed algorithm can reduce encoding time by 38% on aver age compared with HM14.0,with negligible loss of coding efficiency,and it cuts down the computational complexi ty of co ding.