Combining Global and Local Variation for Image Quality Assessment

The information contained in an image is presented by the changes of gray value in spatial domain. Gradient is the basic tool to measure changes, which makes gradient become an important ingredient of most image quality assessment algorithms. However, gradient can only measure local changes, while when human visual system (HVS) perceives an image, it can perceive both local and global changes. Based on this characteristic of HVS, this paper proposes an image quality assessment algorithm by combining global and local variations similarity (GLV-SIM). The algorithm uses Grünwald-Letnikov fractional derivative to measure the global changes and uses gradient magnitude to measure the local changes of the image. Synthesizing the two aspect changes, similarity map between reference image and distorted image is calculated, and then objective score of the image is obtained. Simulation experiments on four databases TID2013, TID2008, CSIQ and LIVE show that, comparing with the algorithm only considering local changes, the proposed algorithm can more accurately simulate the perception process of HVS on image quality and can obtain better consistency between objective scores and subjective scores. Copyright © 2020 Acta Automatica Sinica. All rights reserved.