No-reference Image Quality Assessment Based on Differential Excitation

In order to estimate the degradation of the image distortion level and consider the correlation among pixels, a no-reference image quality assessment algorithm based on differential excitation is proposed in this article. According to the Weber's law, the differential excitation map was obtained and the gradient map of differential excitation was obtained by anisotropy. Then, the differential quantization map was obtained by quantifying differential excitation, and weighted fusion with differential excitation map and gradient map is carried out respectively. Finally, the objective evaluation value of image quality is obtained by support vector regression (SVR) prediction using the acquired features. In LIVE and MLIVE and MDID2013 and MDID2016 databases, the experiment shows that the algorithm is highly robust and low complexity, which can accurately reflect the human image quality of visual perception. Copyright © 2020 Acta Automatica Sinica. All rights reserved.