A Novel Macroblock-Level Rate-Distortion Optimization Scheme for H.264/AVC

摘要

This paper aims at improving subjective quality of the encoded video, and proposes a novel macroblock-level rate-distortion optimization scheme based on perceptual features of the video content for H.264/AVC. In our scheme, three visual distortion sensitivity models are firstly established to minimize the perceptual distortion rather than traditional MAD distortion. During the rate-distortion optimization process, the Lagrange multiplier is adjusted adaptively according to the visual distortion sensitivity of the encoding macrob-locks. Further perceptual tuning of the Lagrange multiplier can effectively restrain distortion in the visually sensitive regions. Experiments have been conducted based on the reference coder JM12.2 of H.264/AVC. Simulation results show the subjective qualities of the decoded frames are improved without compromising PSNR.