基于平均差的Delta CPV快速决策算法
Fast decision algorithm based on mean deviation Delta CPV

为了更精准地呈现深度图的边缘，深度建模模式被引入3D HEVC。深度建模模式中，Delta CPV通过最小失真搜索得到，搜索过程需要遍历所有可能的Delta CPV的集合，这也增加了编码过程的计算复杂度。针对Delta CPV搜索过程计算复杂度高的问题，分析了搜索范围的决策过程，提出利用平均差确定搜索范围的方法。实验结果表明，改进的算法使DMM过程的编码时间减少109%，而BD rate只增加了0.02%。
To represent the edge of the depth map more accurately,the depth modeling mode (DMM) is introduced into 3D high efficiency video coding (3D HEVC).In the depth modeling mode,Delta CPV is obtained by searching for the minimal distortion,and the search process needs to traverse all possible sets of Delta CPV,increasing the computational complexity of the coding process.Based on the high complexity problem of Delta CPV search process,the decision process of the search scope is analyzed,and the method for determining the search range by the mean deviation is proposed.Experimental results show that the improved algorithm reduces the encoding time of DMM process by 10.9% and the BD rate only increase by 0.02%