Abstract —High Efficiency Video Coding (HEVC) exploits quad-tree structured Coding Unit (CU) to improve compression efficiency. It saves about 50% coding bits as compared with the former standard H.264/AVC high profile. However, the computational complexity is dramatically increased for more partition blocks and coding modes supported in HEVC. In this paper, a fast CU decision algorithm is proposed to reduce the number of candidate CUs for HEVC intra coding, which is consisted of two algorithms: a spatial correlation based early CU decision algorithm (SECU) and a Rate Distortion (RD) cost based early CU decision algorithm (RDCU). The depths of spatially neighboring CUs are exploited to skip unnecessary CU size tests first. Then the distribution of RD cost is utilized on the selection of CU sizes. Experimental results demonstrate that the proposed algorithm can achieve the time saving by 46% on average as compared with original encoder with 0.92% BDBR increase and 0.05 dB BDPSNR decrease.
 
Index Terms—HEVC, coding unit, intra coding, computational complexity