An Improvement of Structural Similarity Index for Image Quality Assessment

Abstract

The image quality assessment has been widely used in image processing. Several researches have been developed and carried considering the Human Visual System (HVS). Under the hypothesis that human visual perception is extremely adapted to retrieve structural information from a scene, the SSIM index is the most widely used in this area, which leads to a better correlation with HVS. Despite its robustness the SSIM presents some limitations in the presence of blur affecting images. In this study, we propose an improved version of the SSIM for blur image assessment. The idea is to combine gradient based SSIM score with that of the structural information of the blur. Experimental results show a good performance.

Intensity-Sensitive Similarity Indexes for Image Quality Assessment

“…A. Auxiliary Indexes 1) Sensitivity Index: A sensitivity index sensi is developed to evaluate the improvement in sensitivity of an IQA method (e.g., our new IQA methods) when comparing with a baseline (e.g., the original SSIM), which is expressed by Equation (8). When the sensitivity of a new IQA method is higher than that of the original SSIM, sensi is larger than 0; otherwise, it is smaller than 0.…”

“…To deal with blurred images, a gradientbased structural similarity (G-SSIM) [7] has been developed, which considers the edge of an image as the most important structure information. Following that, several improvements have been made [8], [9] to increase its performance on blurred and noisy images. Additive and spatial pooling methods have also been used to further improve SSIM and G-SSIM [10].…”