Video Quality Assessment Model Based On Blur And Noise Metrics And Applications

With the rapid development and service provision capabilities enhancement of IP and 3G networks, digital video communications has become an important area. The tremendous progress in the video compression and transmission technology provide strong surport for a variety of video business applications. While the digital video technology has brought new business applications and profitable growth for telecom operators, but they are still faced with challenges that the quality of video service for users need to be meet. Therefore, monitoring the quality of video service is needed, and the evaluation for video quality is an essential part of the monitoring function. The purpose of this paper is to study and to realize a video quality evaluation model, which objectively reflect the playback quality of video service fome a user point of view, and provide strong guarantees for the development of network video service.In this paper, by analysising and comparing avariety of video quality evaluation method, a no-reference objective video quality evaluation model is put forward, which can real-time monitor the quality of the online video service. The research content of this paper is that video quality assessment model is based on the metric of blurriness effect and noise effect, and the metric algorithms of blurriness effect and nosie effect was key research. The characteristics of human visual system were considered in the measurement algorithms to make evaluation results more consistent with human visual perception. Experimental results show that the video blurriness indicators, noise indicators and integrated indicators obtained from the model have good correlation with subjective evaluation results. In order that video quality valuation model was applied to fixed and wireless networks, in this paper, the model was implemented as component, and a detailed description of the component design and implementation as well as monitoring applications in the actual network traffic was gived. Experiments show that the component can provide accurate evaluation results of video quality in real monitoring environments, and has a certain application value.