论文部分内容阅读
多视点视频加深度(MVD,multi-view video plus depth)的编码格式包含多个纹理视频序列及其对应的深度图,深度图与对应的纹理视频具有相似的边缘信息。传统的编码模式不考虑两者的联系,单独编码导致复杂度高、编码时间过长。因此,合理利用深度图与纹理图的相似性进行编码,可以有效降低编码复杂度,同时应该确保编码质量不受影响。本文利用深度图的这一特点辅助纹理视频的编码,提出一种帧间快速模式选择算法,充分利用深度图与纹理图之间的相似性,建立一种联合复杂度模型,根据模型得到每个宏块的复杂度。对于复杂度低的宏块,在计算率失真代价之前跳过编码中一些不必要的模式,从而降低编码复杂度。实验表明,本文提出的快速模式选择算法,在保证率失真性能基本不变的情况下,能减少60.57%的编码复杂度,并最高能减少80.64%的编码时间。
The encoding format of MVD (multi-view video plus depth) includes multiple texture video sequences and their corresponding depth maps, which have similar edge information with the corresponding texture video. The traditional encoding mode does not consider the connection between the two, encoding alone leads to high complexity and encoding time is too long. Therefore, it is effective to reduce the coding complexity by reasonably using the similarities between the depth map and the texture map. At the same time, the coding quality should not be affected. In this paper, the feature of depth map is used to assist the encoding of texture video. A fast inter-frame mode selection algorithm is proposed to make full use of the similarity between depth map and texture map to establish a joint complexity model. According to the model, The complexity of the macroblock. For low-complexity macroblocks, some unwanted patterns in coding are skipped before the rate-distortion penalty is calculated, reducing the coding complexity. Experiments show that the fast mode selection algorithm proposed in this paper can reduce the coding complexity by 60.57% and the encoding time by up to 80.64% while guaranteeing the rate distortion performance to be basically unchanged.