An improved R-λ rate control model based on joint spatial-temporal domain information and HVS characteristics

Abstract

With the popularization of smart terminals and multimedia technologies, the video coding standard — H.264/Advanced Video Coding (AVC) and H.265/High Efficiency Video Coding (HEVC) have been unable to meet the needs of various high-definition videos, so the next generation standard —H.266/ Versatile Video Coding (VVC) is under study. In the actual transmission of a video communication channel, rate control plays an important role. However, HEVC rate control based on R-λ model does not adequately take into account the characteristics of the human visual system (HVS). Also, the convergence speed of Least Mean Square (LMS) in HEVC is too slow. In this paper, an improved R-λ(Lambda) rate control model based on joint spatial-temporal domain information and HVS characteristics (IRLRC) is established. In this model, the joint spatial-temporal domain information based on gradient information is used to guide bit allocation for frame and CTU level, where the temporal coefficient is corrected adaptively. What’s more, the Broyden Fletcher Goldfarb Shanno (BFGS) algorithm is introduced, which speeds up the convergence of the proposed model. The experimental results have clearly shown that the proposed IRLRC can achieve better coding performance than HEVC, VVC and other models. In particular, the video sequence based on the proposed IRLRC can meet the needs of HVS and achieve higher optimization for subjective quality.