Research On Rate Control Technology Of HEVC

High-efficiency video coding,as a next-generation video compression standard,introduces a number of advanced coding techniques that significantly increase the compression performance of video.The rate control technology is a particularly important control module in video coding,which guarantees the quality of video coding under limited bandwidth.However,different video applications have different requirements on the rate control technology,in the live video broadcast,the delay of the transmission system is required to be relatively low,while in the medical video,the video quality is required to be high.For different requirements above,this paper optimizes the rate control algorithm in different forms.1.In this paper,aiming at the shortcomings of the rate control algorithm in the-0103 KJCTVC proposal in low-delay applications,the bit allocation method in the algorithm is improved.First of all,the improved algorithm in this paper synthetically considers the buffer state associated with low delay to allocate bits to the frame layer;then,at the levelLCU-,it is found through experiments that the number of bits actually consumed by encoding is linearly related to the LCU complexity calculated from the gradient,and then the number of bits allocated to the uncoded LCU according to this complexity;finally,coding parameters are obtained according to the R-? model.The results show that the improved algorithm in this paper makes the code rate fluctuation less,the state of the buffer area is more stable,and the video low-delay transmission is realized.2.In order to obtain high-quality video suitable for human visual system under limited bandwidth,this paper presents a code rate control algorithm based on the region of interest of the human eye.First of all,in order to make the video quality of different regions smoothly transition,the region of interest is preprocessed;then,we get the pixel-level region of interest map(ROI map)as the bit allocation weight,and use the frame-level and levelLCU-ROI map to allocate the number of bits to the frame layer and LCU layer respectively;finally,the quality of the region of interest is controlled by adjusting the pixel weight value k.The results show that the algorithm presented in this paper significantly improves the rate-distortion performance and subjective quality of the region of interest when the overall frame quality PSNR does not change much.