Research On Hight Compression Rate And Low Complexity Coding Based On H.265/HEVC

With the continuous development of video applications to HD and UHD,the proliferation of video data continues to challenge the video transmission and storage capacity under the limited channel bandwidth and storage capacity,which makes the widely used H.264/AVC video coding standard gradually unable to meet the needs of efficient video storage and transmission.Therefore,the Joint Collaborative Team on Video Coding released the new generation of international video coding standard H.265/HEVC in 2013.Although H.265/HEVC has achieved another leap in video compression efficiency by virtue of its advanced coding tools,the contradiction between coding performance and coding complexity still exists on the way to large-scale popularization of H.265/HEVC standard.Based on the third generation of efficient video coding standard H.265/HEVC,this dissertation tries to make a new breakthrough in the theoretical research of high compression rate video coding and low complexity video coding.The main research contents of this dissertation are as follows:1.A scene cut detection algorithm based on local luminance histogram is proposed.Based on the excellent representation ability of local luminance histogram for inter frame content change,and aiming at the efficient detection of scene cut in video,the inter frame content change metric and scene cut detection strategy based on scene cut detection window are designed.The proposed scene cut detection algorithm based on local luminance histogram can achieve efficient scene cut detection at ultra-low computing cost,and improve the accuracy of scene cut detection and video compression efficiency.2.An adaptive slice type decision algorithm based on local luminance histogram is proposed.Based on the good representation ability of local luminance histogram for video temporal characteristics and the efficient adaptive decision-making of slice types in video,a better compatible fixed slice type decision-making algorithm and an adaptive slice type decision-making process based on the overall difference between adjacent frames are designed.The proposed adaptive slice type decision algorithm based on local luminance histogram can realize the efficient adaptive slice type decision without any extra computation cost,and effectively improve the efficiency of video compression.3.An adaptive quadtree depth prediction mechanism is proposed.Based on the temporal and spatial correlation between the depth of coding unit division and the quadtree division method in H.265/HEVC standard,this dissertation designs the prediction algorithm of similar region flag,similar region depth range and non similar region depth range.The proposed adaptive quadtree depth prediction mechanism can significantly accelerate the coding process while effectively ensuring the coding quality.