A novel macroblock-level rate-distortion optimization scheme for H.264/AVC

Abstract

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 macroblocks. 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. © 2008 Springer Berlin Heidelberg.