基于模式复杂度的深度视频快速宏块模式选择算法

Depth video fast macroblock mode selection algorithm based on mode complexity

为了降低深度视频编码的复杂度,提高深度视频编码的效率,本文利用深度视频纹理复杂度的一致性,提出了一种深度视频快速宏块模式选择算法.该算法首先对不同区域模式复杂度进行分析,定义一个宏块模式复杂度因子,根据该因子将深度视频划分为简单模式区域和复杂模式区域.然后,分析深度视频的宏块模式分布情况及宏块模式选择的相关特性,为提出的算法提供理论依据.最后,根据模式区域选择相应的编码策略完成深度视频宏块模式的快速选择.实验结果表明,提出的算法在几乎不影响编码率失真性能及保证绘制质量的情况下可明显降低计算复杂度,能节约65.57％～92.72％的编码时间;而在低码率的情况下节约的编码时间会更多.因此,该算法在较低网络带宽限制下更加有效.

To reduce the complexity of encoding depth video and to improve the efficiency of encoding depth video,a depth video fast encoding algorithm was proposed by utilizing the consistency of the complexity of depth video texture.Firstly,the mode complexity in different areas was analyzed.The complexity factor of a macroblock mode was defined,and the depth video was divided into two kinds of areas,simple area and complex area,by using the factor.Then,the distribution of macroblock mode of depth video and the correlation characteristics of macroblock mode selection were analyzed in detail to provide a theoretical basis for the proposed algorithm.Finally,a fast selection algorithm for depth video macroblock mode was implemented based on different encoding strategies from the simple area or the complex area.Experimental results show that the algorithm can save 65.57％-92.72％ encoding time while maintaining the encoding rate-distortion performance and the quality of virtual view.Furthermore,more encoding time can be saved in the condition of low bit rate.Therefore,this algorithm is more effective under low network bandwidth limitations.