纹理-深度图共同优化的3D-HEVC帧内快速算法

A Fast Algorithm for Texture-Depth Co-optimization in 3D-HEVC Intra Prediction

随着立体及3D视频需求的日益增多,针对3D视频编码方法的研究受到越来越多的关注。3D-HEVC编码标准对采用纹理和深度图格式融合的3D视频进行编码,由于加入了深度图编码,因此新增了深度图编码模式、组件间预测和分段直流编码等技术,使其编码复杂度急剧升高。为了减少3D-HEVC的编码时间,本文提出了针对纹理图和深度图的编码单元(Coding Unit,CU)尺寸提前决策快速算法。利用梯度矩阵和作为当前CU和子CU复杂度的判断依据,将CU分为三类:不划分CU(Non-Split Coding Unit,NSCU)、直接划分CU(Split Coding Unit,SCU)以及普通CU。对NSCU,跳过小尺寸的帧内预测过程;对SCU,直接跳过当前CU的帧内预测过程;对普通CU,执行原平台操作。实验结果表明,与原始平台相比,本文算法在合成视点质量基本不变的情况下,平均减少40.92%的编码时间;与最新的联合纹理-深度图优化的3D-HEVC快速算法相比,可以在质量相当的情况下减少更多的编码时间。

With the increasing demand for stereo and 3D video,the research on 3D video coding has attracted more and more attentions.3D-HEVC standard is for 3D video coding with texture and depth map format.Due to the addition of depth map coding techniques,the coding complexity increases dramatically.Therefore,a fast intra coding algorithm for both texture and depth maps in 3D-HEVC was proposed.First,CU was divided into three categories:Non-split Coding Unit(NSCU),Split Coding Unit(SCU),and ordinary CU,by using the Sum of gradient matrix(SGM)as the complexity judgment basis of current CU and Sub-CU.For NSCU,the small size of the intra-frame prediction process was skipped;for SCU,the intra-frame prediction process of the current CU was skipped directly;and for ordinary CU,the original platform operation was performed.Experimental results show that compared with the original platform,the proposed algorithm reduces the coding time by 40.92%on average when the quality of synthetic viewpoints is basically unchanged.Compared with the state-of-the-art texture-depth co-optimization fast algorithms for 3D-HEVC,the outperformance of time saving is achieved with relatively equivalent quality.