Research On Fast Intra Coding Module Based On HEVC

High Efficiency Video Coding(HEVC),as the most widely used Video Coding standard of the new generation,can significantly improve the efficiency of video compression compared with the Advanced Video Coding(AVC)standard of the previous generation.However,the computational complexity also increases exponentially,which brings great challenges to the hardware design of encoders.Therefore,the optimization of algorithms,especially the intra-frame coding algorithm,has become a hot research topic in recent years.Based on the texture information of image blocks,this paper optimizes the algorithm from the two aspects of the division of coding units and the decision of prediction mode,proposes a fast algorithm of intra-frame coding to reduce the complexity of intra-frame coding,and realizes the design and verification of its hardware structure.The main research contents are as follows:Firstly,the partition algorithm of coding unit is optimized.In this paper,firstly,the gradient information of the current coding unit is extracted to represent its texture complexity and the differences among subblocks.Then,the empirical thresholds of different sizes are set according to the coding units of different sizes,and the gradient and difference values among subblocks are used to decide whether the current coding unit should continue quadtree partitioning or terminate the partitioning process in advance.Thus,the ergodic complexity of coding unit depth is reduced,and it is convenient to implement in hardware.Secondly,the decision algorithm of the prediction model is optimized.In this paper,the gradient direction of the current prediction unit is used to reduce the number of candidate model sets,and the projection value information of the gradient value on each prediction model is used to determine the optimal Angle prediction model,and then the texture complexity information of the current prediction unit is used to determine the optimal Angle prediction model in the plane model,the DC model and the optimal Angle model.Thus,the computational complexity of prediction mode is reduced and it is convenient to implement in hardware.Then,the above two optimized algorithms are combined and the performance of the algorithm is verified.Compared with the standard algorithm in the standard coding software(HEVC test Model,HM),the in-frame coding optimization algorithm proposed in this paper can reduce the coding time by 50.35%,increase the bit rate by 1.78% and reduce the peak signal-to-noise ratio by 0.074 db.Therefore,under the premise of less loss of image coding quality,The coding time is greatly reduced,and the computational complexity is effectively reduced.Finally,the hardware design scheme of the in-frame coding module is implemented and its function is verified.The hardware implementation scheme design and function verification are carried out on the three modules of the division module of the coding unit,the decision module of the prediction mode and the intra-frame prediction module respectively.Finally,the functional correctness of the hardware design of the intra-frame coding module is proved by the way of hardware and software cross-verification.