Research On Fast Coding Algorithm For HEVC And 3D-HEVC

As the improvement of video display technology,high-definition(HD)video and Ultra High Definition Video(UHDV)have progressively became the mainstream of video.However,it always with a large amount of video data increments which hinder the video storage and transmission in internet.The present video coding standard H.264/AVC cannot fulfill the compression efficiency request of HD video or UHDV.In this case,the new video coding standard High Efficiency Video Coding(HEVC)arises at the historic moment.Apart from that,as the progress of 3D technology there are some new display technique appeared,such as the multi-view display technique and the free viewpoint display technique and so on.Those new technique require more views to be coded and transmitted which will bring higher video data.Thus,the 3D video coding also become a hot research point and the 3D-HEVC is proposed as the 3D extension of HEVC.HEVC brings two times compression efficiency for H.264/AVC under the same video quality through many new coding tools.But those new coding tools bring a high computational complexity,in addition the 3D-HEVC introduce the depth map coding technology which bring more complexity augment.The high computational complexity brings an obstacle for the application of HEVC and 3D-HEVC.Thus,it is very important to optimize the coding algorithm to reduce the coding complexity for HEVC and 3D-HEVC.In this paper,we introduce the basic knowledge of video coding and its 3D extension,and then summarize the history of video coding standard briefly.We also analysis the structure of HEVC prediction coding,and introduce the new coding tools which import high computational complexity for video coding.After that we enumerate some hot research points of fast video coding method.And then we also propose three fast coding methods to reduce the high computational complexity of the process of CU size and prediction mode detection for HEVC and 3D-HEVC prediction coding,those three methods can be seen as follow:1.A fast Inter-frame encoding scheme using the edge information and the spatiotemporal encoding parameters to reduce the encoder complexity of HEVC was proposed,which consists of a fast all 2N × 2N modes decision method,a fast CU depth level decision method and a fast PU mode decision method.This method uses edge information to reflect the structure complexity and uses the difference of edge information between current CU and its spatiotemporal CUs to express the edge similarity and the edge movement.And then,associate the edge similarity and the edge movement with CU depth levels and PU mode RD costs of spatiotemporal adjacent CUs to fulfill the early termination of CU split and the PU mode selection.There are 53.7% and 54.9% encoding time savings on average for various test sequences under random access and low delay B conditions,respectively.2.A fast coding algorithm for HEVC based on video contents was proposed.We statistically analysis the features of video contents from three aspects,the pixel gradient,the block mean value,and the block variance of CU.And then,associate these features with the coding information of spatiotemporal adjacent CUs to realize fast CU size decision and fast prediction mode decision.The experimental results show that the proposed algorithm can save 58.9 % and 57.6% computational complexity on average under random access and low delay B,respectively.3.We analyzed the multi-view extension of HEVC and proposed a method based on the contents of depth maps to reduce the encoding computational complexity of texture videos for 3D-HEVC.The numbers of different depth values and the mean square deviation of depth values in one CU are used as parameters to realize early CU size decision and early Skip mode decision.At last we conclude our works in this paper and predict the hot research area of video coding in the future.