一种基于时空相关性的编码单元深度快速分级判决算法

A rapid classification decision algorithm on CU depth based on temporal-spatial correlation

新一代视频编码标准HEVC采用编码树单元(CTU)结构进行编码,通过遍历比较所有不同深度编码单元(CU)的率失真代价值得到最佳编码单元,在显著提升编码效率的同时,计算复杂度增加数倍。为此,本文提出了一种基于时空相关的编码单元深度快速分级判决算法,通过量化分析时空相邻CTU/CU之间相关性权重,利用已编码时空相邻CTU/CU的最佳深度预测当前CTU/CU可能的深度范围(DR)和深度值,跳过和提前终止不必要的深度计算。仿真结果表明,在不同的编码配置条件下,本文算法在保证编码性能的同时,平均可节省40%以上的编码时间,极大地降低了计算复杂度。

The emerging high efficiency video cod ing （HEVC） standard uses the structure of coding tree unit （CTU） to compress vid eos.Rate-distortion costs of all the different depths from depth 0to dep th 3are compared transversally to get the optimal coding unit （CU）.The encoding efficie ncy is improved significantly by these new added technologies in HEVC.However,greater computation complexity is i ntroduced compared with previous standards,which could limit its practical application.Thus,it is of vital importance for HEVC to increase the encoding speed.A rapid classification decision algorithm on CU dept h based on temporal-spatial correlation is proposed in this paper to solve this problem.First of all,correlation wei ghts of temporal-spatial adjacent CTUs/CUs are analyzed.Then,the possible depth ranges and depth values of current CTU/CU can be predicted in advance via depths of these coded CTUs/CUs.As a result,unnecessary CU size decis ions can be skipped and terminated.According to the simulation results,conclusions are obtained that enc oding complexity on average can be reduced by 40% under condition of different encoding configurations,while ma intaining almost the same RD performance as the original HEVC encoder.Our proposed algorithm reduces the cod ing complexity greatly compared with the original HEVC encoder.