A Novel Algorithm for Video Smoothness Evaluation

摘要

By making use of the temporal correlations of video sequences, a novel algorithm for video smoothness evaluation is proposed, which reflects the impact of frame loss on perceived visual quality. This metric particularly aims at measuring the perpetual quality degradation caused by video temporal fluctuations, based on the phenomenon that human visual system(HVS) pays more attention on difference among frames rather than a signal frame's sharpness. Then a visual smoothness degradation evaluation model is established, in which both source smoothness degradation and receiver smoothness degradation are measured to evaluate perceived video smoothness. Experimental results and subjective assessment results demonstrated the high performance of this algorithm on prediction of perceptual smoothness degradation.