Research On Bit Allocation Algorithm For Versatile Video Coding

In recent years,the rapid development of computer and media technologies has led to significant progress in industries related to video,and the video encoding industry is constantly introducing new encoding standards to meet the widespread demand for video compression in the industry.The Versatile Video Coding(VVC)standard has introduced various advanced encoding techniques,which comprehensively improve the compression performance of encoders.In video communication,rate control is an essential control module that can ensure video quality under bandwidth limitations and is currently one of the research hotspots.The VVC standard adopts advanced rate control algorithms,which have improved rate control accuracy and video quality,but its frame-level bit allocation scheme does not consider the impact of subjective perception on human vision,and its Coding Tree Units(CTU)level bit allocation scheme does not fully consider the spatial and temporal complexity of the image,these shortcomings leading to suboptimal bit allocation effects.This paper addresses these two issues and proposes corresponding optimization algorithms.Firstly,this paper proposes a perception-based frame-level bit allocation algorithm.Firstly,a video fusion saliency map that conforms to the VVC encoding characteristics is extracted based on the perceptual characteristics of the human visual system,using motion saliency and contrast saliency.Secondly,a perceptual weight reflecting the degree of human attention to video frames is designed based on the fusion saliency map.Finally,the perceptual weight is used to control the bit allocation process.Experimental results show that the optimization algorithm can allocate more bits to video frames with high visual attention,effectively improving the visual experience.Secondly,this paper proposes a CTU-level bit allocation algorithm based on spatial and temporal complexity.Firstly,a complexity measurement factor is constructed based on image information entropy and mean absolute deviation to characterize the image complexity of CTUs in spatial and temporal dimensions,which is used as the bit allocation weight at the CTU level.Secondly,a central enhancement operation is designed based on human viewing habits to enhance the bit allocation weight of CTUs located in the center of the picture and reduce the bit allocation weight of CTUs located at the edge of the picture.Experimental results show that the optimization algorithm can effectively improve the objective quality of the video.