Video Quality Metric for Consistent Visual Quality Control in Video Coding

The visual quality consistency is one of the most important issues in video quality assessment (VQA). When people view a sequential video, they may have an unpleasant perceptual experience if the visual quality of video frames is inconsistent even though the average visual quality of the video is not too bad. Thus, the consistent visual quality control is mostly expected in real-time video communication. Additionally, in conventional video communication, the channel bandwidth and buffer resources are limited. The unfair distribution of encoding resources among video frames would result in not only inconsistent visual quality but also other types of spatial distortions. In this paper, a new objective visual quality metric (VQM) is firstly proposed for measuring the video quality in video coding. It makes full use of the information of video coding without extra computational complexity. Secondly, a visual quality control algorithm is proposed to ensure the consistent visual quality of video coding under the given channel and buffer resources. Finally, the experimental results indicate that the proposed VQM is consistent well with the human visual system (HVS). In addition, the consistent visual quality, better rate-distortion efficiency, accurate bit control and compliant buffer can be achieved by the proposed visual quality control algorithm.