3D-HEVC深度图帧内模式快速决策算法

Fast decision algorithm for intra mode in depth map of 3D-HEVC

作为高效视频编码(HEVC)扩展之一的3D-HEVC在合成视点上有着良好的效果,但为保证深度图的编码质量,所加入的深度建模模式(DMMs)使其具有了极高的编码复杂度。为了降低深度图的编码时间,本文提出一种基于边缘检测的深度图帧内模式快速决策算法,首先使用Roberts算子进行边缘检测,判断是否需要遍历深度建模模式;然后根据普通模式的粗选结果和最有可能的模式(MPM),采取给预测单元(PU)和角度模式分类的方法,大幅降低楔形模式的遍历数量,从而大幅减少了深度图的编码时间。实验结果表明,在全帧内模式下,与HTM-16.2相比,本文所提算法可节省35.4%的深度图编码时间,编码比特率只增加了1.06%,编码效率的损失可以忽略不计。对于具有边缘数量较多且明显的序列,编码时间减少的也较多。因此,本研究对3D-HEVC的实际应用具有一定的参考价值。

As an extension of High Efficient Video Coding(HEVC),3 D-HEVC has a good effect on the synthesis point of view,but in order to ensure the coding quality of depth map,the added Depth modeling modes(DMMs)makes it has extremely high coding complexity.In this paper,in order to reduce the encoding time of depth map,a fast decision algorithm for intra mode of depth map based on edge detection is proposed.Firstly,the Roberts operator is used for edge detection to determine whether the DMMs mode needs to be traversed.Then,according to the Rough Mode Decision(RMD)results of the common model selection and the Most Probable Mode(MPM),the traversal number of wedgelet patterns is greatly reduced by divide prediction unit(PU)and angle patterns into a few categories,thus reduced much depth map encoding time.The experimental results show that,compared with HTM-16.2,the proposed algorithm saves 35.4%of the encoding time of depth map and only increases 1.06%of BDBR in the All-intra mode,and the loss of encoding efficiency can be ignored,the coding time reduction is particularly high for much steep edges sequences.Therefore,this study shows a certain reference value for the real-time application of 3 D-HEVC.