Research On VVC Rate Control Algorithm Based On Visual Perception

Nowadays,In the era of the internet,video information has permeated every aspect of people’s lives,from watching videos on mobile devices to national defense monitoring.Videos make production and life more convenient and efficient.However,in the new generation video coding standard VVC framework,there is no separate rate control algorithm for human visual perception,but traditional algorithms are directly followed.Traditional algorithms do not consider the visual characteristics of the human eye,and the direction of bit rate allocation does not match the real perception of the human eye.In response to these issues,this article designs a visual perception factor based on the Human Visual System and applies it to two rate control optimization algorithms.The specific research and innovation content is as follows:In response to the lack of indicators that fully reflect the characteristics of human visual perception in the coding framework,and the lack of consideration of human visual perception characteristics in existing Rate Distortion Optimization technologies,this paper proposes a rate distortion optimization algorithm based on human visual indicators,which leads to the inconsistency between the coding results and the subjective perception of the human eye.By integrating the brightness characteristics,frequency domain characteristics,and time domain characteristics of the human visual system,a human visual factor has been designed,which can effectively determine the human visual sensitivity and visual importance of the coding tree unit region.Using this visual factor in the rate distortion optimization model,the Lagrange multiplier can be adjusted according to the different visual characteristics of the coding unit in each frame to guide the optimization of the rate distortion process.The experimental results show that the proposed algorithm achieves an average BD-Rate gain of 0.38% based on VMAF metrics,saving an average of 1.66% of BDBR under the same quality,while only increasing time by 2.3%.This paper proposes an adaptive QP offset selection algorithm based on visual characteristics to address the mismatch between the allocation direction and visual sensitivity of the QP quantization algorithm,which does not link CTU with human visual factors during rate allocation.Utilizing the human visual factor and the minimum perceptible distortion value JND proposed in the rate distortion optimization algorithm,which can reflect human eye perception,to jointly guide QP offset selection.Firstly,the K-Means clustering method is used to classify the visual levels of CTUs based on visual indicators from different angles.Then,the SVM multi classification model is used to determine the specific segmentation threshold.Finally,the corresponding QP offset values are selected for different levels of CTUs,allocating more bit rates to areas that are more interesting to the human eye and easier to detect distortion,and improving the visual distortion problem caused by compression.The experimental results show that the adaptive QP offset selection algorithm achieves an average BD-Rate gain of 0.91% based on VMAF indicators,and also achieves an improvement of about 5 points in VMAF scores.Compared with traditional rate control algorithms,this paper designs a human visual perception factor V.The rate distortion optimization algorithm based on visual perception factor can save more bit rates and improve perceptual encoding performance.The quantization optimization algorithm guided by visual characteristics makes the rate allocation more in line with the subjective perception of the human eye,improves the visual distortion problem in compression,and has minimal time loss,proving that the optimization algorithm proposed in this paper has practical application value.