| High Efficiency Video Coding (HEVC) is a next generation standard that isdeveloped jointly by ITU-T and ISO/IEC which regarded as the most famous videocoding standard developers. With the increasingly high quality requirements of videocoding, H.264has been unable to meet the needs of people. But compared withH.264, HEVC enhance coding efficiency when also brings some difficult problems,high computational complexity increases significantly. How to reduce the complexityof HEVC encoding algorithm based on ensuring the coding performance, become animportant issue in the research of HEVC standard.In order to improve the coding performance, this paper firstly introduces somebasic concepts of HEVC coding standard, and then studied the inter mode predictionand CU partition algorithm. In this paper, on the two part of the contents of PU and CUrespectively from the point of view put forward two kinds of fast algorithms.The firstexercise first feature of video sequence statistics,appropriate experimental valuesobtained,the correlation of PU modes between the surrounding CUs of thecorresponding located CUs in the previous frame and the current CU is analyzed,adivision of PU-based domain correlation fast algorithm is proposed.Compared with thestandard algorithm in HM12.0test model, in the case of I frames, in which PSNR isalmost unchanged, reduce encoding time by54.09%.Second quantitative parametersconsidering the texture characteristics and the encoded video frame in the effectoptimal coding unit (CU) mode selection, by dividing these factors to early terminationCU. The experimental results show that compared with the original HM-12.0algorithm,in all I frame and PSNR unchanged, the encoding time is reduced by37.42%, thecoding speed is significantly improved. |
