The Research On Fast CU Partition And Intra Mode Decision Algorithm For HEVC

With the entry of digital multi-media era and the growing of Internet Technology,every day,people produce a huge amount of digital image information,and the video image information we face is generally compressed.Video coding(compression)is a key part of multimedia technology designed to reduce data redundancy and simplify the amount of information.In the era of highly prosperous information,visual-related businesses are becoming mainstream.Digital media TV,distance education and surveillance,short video and live broadcasts are gradually integrated into people's lives,changing people's living habits and production methods.The degree of digitization is getting higher and higher,and the network infrastructure is constantly being upgraded and improved.High-definition video and diversified application scenarios are becoming more and more popular.In order to solve the growing demand for high-performance coding,a new generation of video coding standards(HEVC)gradually replaced the H.264 standard and went to the center of the video coding arena.In order to accelerate the coding time,improve the performance of the HEVC standard,and save bandwidth,this paper proposes a fast algorithm for CU partitioning and pattern prediction of HEVC coding standard.This paper first introduces the basic knowledge of HEVC,coding framework,rate distortion technology,and analyzes the CU partitioning and pattern prediction process.The current rate-distortion optimization technique can not consider the shortcomings of subjective evaluation well;in the CU partitioning of intra prediction coding,the existing algorithms do not fully consider the information in the spatial domain;in the PU block mode decision,the complexity of HEVC rate distortion calculation The degree is still very high,and there is more room for optimization.In response to the above problems,this paper has carried out the following research:First of all,this paper briefly introduces the development process of HEVC and summarizes the current research status,and then introduces the system architecture and related knowledge of HEVC.Then,the research on HEVC rate distortion optimization technology,CU partitioning and pattern prediction process in intra prediction coding is carried out,which lays a theoretical foundation for the proposed algorithm.Secondly,this paper first analyzes the rate-distortion optimization algorithm in HEVC,and proposes a rate-distortion optimization algorithm based on visual characteristics by changing the distortion measure.Then,based on the analysis of the relationship between CU block segmentation and intra-predicted image texture values,a fast CU segmentation algorithm based on image texture information is proposed.In the process of intra prediction mode selection,by the statistical analysis of the prediction mode in HEVC coding,the idea of classification processing is proposed,and then the candidate mode for entering the RDO process is reduced according to the SATD order,thereby reducing the number of prediction modes.At the end,the performance of the proposed algorithm is tested experimentally.In the fullframe(AI)mode test,experimental results indicate that compared with the algorithm in the HM16.16 encoder,when the coded image quality of the CU partitioning and mode decision fast algorithm is almost unchanged,the average coding time is saved.52.7%,and the coding BD-rate only increased by 0.61%,effectively improving the coding efficiency.