Objective method for assessment of video quality using wavelets

Abstract

In this paper, we propose a new objective method for assessment of video quality. The method is based on a recursive biorthogonal wavelet transform and is the double-ended objective video quality assessment method capable of receiving as input a processed video sequence and its corresponding source video sequence. We apply 2D wavelet transform to frames of video sequences. After the wavelet decomposition, each frame is divided into several subbands corresponding to spatial frequencies and this would make it possible to take into account the sensitivity of human visual system to stimuli of different spatial frequencies. Then we obtain an objective quality score by computing PSNR from weighted sum of the squared errors of subbands.