An adaptive visual quality optimization method for Internet video applications

A substantial proportion of current Internet videos are poor in quality. To improve the visual experience, many visual quality optimization algorithms, such as denoising, sharpening are usually used. However, the unsuitable denoising will blur the image, the oversharpening will result in overshoot artifacts. To resolve this problem, many methods have been proposed to adaptively adjust the parameters based on the content of the videos. However, these methods are characterized with high computation complexity and are not easy for realtime video applications. In this paper, an effective visual optimization method, both denoising and sharpening, for low quality Internet videos is presented. A technique is used to get the approximate visual shape map from the video sequences. The shape information is exploited to adjust the denoising strength and sharpening masks. Thus a content adaptive visual optimization algorithm is achieved. The experimental results show that the proposed algorithm has good performance and can be used for Internet video applications.