IMAGE SUPER-RESOLUTION USING NONLOCALLY CENTRALIZED SPARSE REPRESENTATION AND FIELDS OF EXPERTS PRIORS

Although the high-resolution image can be reconstructed in the super-resolution reconstruction algorithm based on nonlocally centralized sparse representation, it does not effectively suppress the noise in the reconstructed image. It is necessary to select the appropriate priors constraint to reduce the impact of noise in the reconstructed image, so that the quality of reconstructed image can be improved when observed image have different noises. In this paper, we consequently introduced the Fields of Experts model into the nonlocally centralized sparse representation framework as the prior knowledge of image, which can effectively enhances the effect of image denoising. Moreover, the experimental results show that the proposed method can reduces the noise and gets better reconstructed images.