Full-Reference Image Quality Assessment Approach Based on Image Separation

In this paper, a full-reference image quality assessment (IQA) approach based on blind source separation algorithm is present. After the reference image is divided into several patches, the scatter plot of these patches is contained within a simplicial cone. The mixing matrix is estimated by the parameter of the simplicial cone. The sources of the reference image and the distorted image are extracted by mutiplying the image pathces by the mixing matrix. Different sources of the reference image and the distorted image are used to calculate the structural similarity (SSIM) index values. SSIM values of different sources are weighted and summed to calculate the final objective IQA criterion. Experimental results show that the IQA approach outperforms the traditional single source IQA approach.