Image quality measurement through structural similarity based on higher order moments
Swarnjeet Kaur, Harpreet Kour, Debashis Sen · 2016
Structural similarity index (SSIM) considers the loss of structural information as a degradation of quality. Structure of original and distorted images are compared by using low order moments, which are mean, variance and correlation. In this paper, we extend the SSIM by incorporating shape parameters of distributions, which are the higher order moments based skewness and kurtosis. We show that skewness and kurtosis adds useful extra information to SSIM, which is relevant in quantification of local structures. We also show that this additional information improves the correspondence of SSIM with human perception. Results are taken on various types of distorted images of a standard dataset and new SSIM is validated against SSIM index, mean square error and subjective ratings.