Font Size: a A A

Research On The Related Techniques Of High Efficiency Video Coding

Posted on:2019-02-03Degree:MasterType:Thesis
Country:ChinaCandidate:P RanFull Text:PDF
GTID:2348330569487835Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
With the gradual diversification of video applications and the continued popularity of ultra/high-definition(HD/ UHD)video,video data has increased exponentially.Because the growth rate of storage and channel resources is always slower than the growth rate of video data,it is urgent to improve the performance of video compression,so as to cope with the growth of video data caused by HD/ UHD application and the diversification of video applications.Against this background,the two major international video coding standardization organizations MPEG(Moving Picture Experts Group)and VCEG(Video Coding Experts Group)have jointly established a new international standard,so-named High Efficiency Video Coding(HEVC).HEVC still uses the hybrid coding structure,and introduces the flexible block partition technology,multi-angle intra prediction mode,Merge mode and some other new technologies.HEVC finally improve the compression performance relative to H.264/AVC(Advanced Video Coding)—in the range of 50% bit-rate reduction for equal perceptual video quality.Since the HEVC standard was officially released in 2013,the research work on its related technologies has been continuing.In order to further improve the coding efficiency,we do some reseach on the related technologies of the video coding standard HEVC.The primary work and innovation of the thesis are as follows:1.Aiming at the limitation of translation motion model to complex motion description,we analyze the relationship between moving forward visual video and perspective model,and propose an inter prediction method for moving forward visual video based on perspective model.In order to avoid the solution and coding of too many model parameters,this method is implemented by the way of template traversal.The proposed method is integrated into the HEVC reference software HM16.7.Compared with the original HM16.7 encoder,the proposed method is proved to be effective.2.We do some research work on the Merge mode of HEVC,and analyze the shortcomings of the Merge mode in the process of inter prediction.According to the shortcomings,we propose two Merge mode combination weighted prediction methods.One is based on direction,and the other is based on distance.The proposed two inter Merge mode combination weighted prediction methods are integrated into the HM16.7.Compared with the original HM16.7 encoder,in low delay P configuration,the proposed method(based on direction)can bring 1.39% and 1.45% BD-rate savings on average for test sequences of Class C and Class D,respectively.In the same test condition,the proposed method(based on distance)can bring 2.94% BD-rate savings on average for test sequences of HEVC.The experimental results show that the proposed two inter Merge mode combination weighted prediction methods can improve the coding performance significantly.
Keywords/Search Tags:video coding, inter prediction, moving forward visual video, perspective model, Merge mode
PDF Full Text Request
Related items