The Research On Mode Optimization For High Efficiency Video Coding

High Efficiency Video Coding(HEVC)is a new video coding standard.Compared with the H.264/AVC standard,the coding efficiency of HEVC standard has been improved significantly.But at the same time,the encoding complexity of HEVC is increased by several times.The high encoding complexity seriously hinders practical applications.Therefore,this paper focuses on the coding optimization method of HEVC standard,in order to reduce the coding complexity as well as maintain the excellent video coding quality.Main contributions in this thesis include:1.The partition decision process of coding unit(CU)and the mode decision process of prediction unit(PU)in HEVC have been analyzed in detail.Based on it,the coding strategies are optimized.2.Aiming at the high complexity of HEVC intra coding,a low-complexity encoding method for HEVC based on adaptive learning random sample set is proposed.The low-complexity coding method for HEVC includes two parts: a learning encoding method and a fast encoding method.Each video sequence is grouped according to its frame rate.The first two frames of each sequence are learning encoding frames,where the thresholds of CU classification and the thresholds of rate distortion cost have been obtained.The subsequent frames are fast encoding frames,where CU partition and PU mode decision have been optimized based on the previous thresholds.The low-complexity encoding method for HEVC can reduce the encoding time by 34.70% on average with neglecting degradation of encoding performance.3.Aiming at the high complexity of HEVC screen content coding,a fast coding method based on neighboring-correlation for HEVC screen content coding is proposed.Three parts are included: a fast CU partitioning algorithm based on neighboring-correlation and a fast mode selection decision algorithm based on neighboring-correlation and a fast CU partitioning algorithm based on coding bits.Experimental result shows that the proposed method achieves the average time-saving of52.18% with 1.42% Bj?ntegaard difference bitrate(BDBR)increase and 0.17 dB Bj?ntegaard difference PSNR(BDPSNR)decrease on the HEVC screen content coding test model(SCM)2.0.And the improved method achieves the average time-saving of 54.57% with 1.94% BDBR increase and 0.28 dB BDPSNR decrease on the HEVC screen content coding test model(SCM)4.0.