A new SVD-based image quality assessment

In recent years, many efforts have been performed in order to design an algorithm assessing perceptual image quality based on human visual system. Although some impressive metrics have been presented, full reference image quality assessment (IQA) is still a challenging issue. In this paper, we present a new SVD-based IQA method in which the structural similarity between the reference and distorted image is utilized as a key factor for measuring the imposed distortions. The experimental results show that the proposed algorithm can effectively evaluated the image quality in a consistent manner with human visual perception.