Background-based Objective Measure for Predicting Subjective Visual Quality of Distorted Image

摘要

This paper proposes a new image objective quality measure, which is based on background activities. The proposed background-based quality measure is base on singular value decomposition (SVD) of block and background images. The proposed objective quality measure with weighted background activities and singular vectors error between the corresponding original and the distorted image blocks exhibits a better subjective visual quality prediction. Experiments results also show that the proposed background-based SVD image quality measure truly provides a better correlation with the human eyes perception than the existing measures