Fast video enhancement algorithm using real-time image restoration framework

摘要

A fast video enhancement algorithm is proposed. The algorithm enhances the quality of video by removing degradation due to subsampling and compression. In order to incorporate a priori knowledge into the iterative minimization process, a real-time framework of iterative restoration is devised. Real-time implementation is realized by segmenting moving and motionless regions, and regularized iterative restoration and simple filtering are separately applied to the different regions. The assumption of motionless region holds for wide range of low-resolution, fixed surveillance cameras and video telephony.