应用多恰可感知失真等级的视频感知编码

To explore the perceptual redundancy in video, and to improve the subjective quality of reconstructed videos, a perceptual model for video distortion assessment based on multi-just-noticeable-distortion level was proposed. Then the model was applied to high efficiency video coding (HEVC) to implement perceptual coding, where the proposed model was used to classify the video content according to its perceptual effects. Rate allocation of the traditional encoder was improved, where intra and inter frames were processed respectively, and quantization parameters (QP) was updated according to the bits reallocated. The experimental results show that compared with the HEVC test model (HM), the proposed algorithm achieves better perceptual quality at a similar coding bitrate compared with the original HM.