Image quality assessment based on multi-channel regional mutual information

In this paper, a novel method based on multi-channel regional mutual information (RMI) is proposed to assess image quality. In the proposed method, the wavelet transform is firstly used to decompose the image into different frequency subbands to calculate RMI values. Then multi-channel RMI is obtained by weighted sum of RMI values in the various wavelet frequency subbands. The performance of the proposed algorithm is compared with that of such universal assessment methods as peak of signal-to-noise ratio (PSNR) and structure similarity (SSIM). Experimental results demonstrate that the proposed method is highly effective for evaluating image quality and it outperforms the assessment methods based on PSNR and SSIM.