Blind Image Deblurring Using Patch-Wise Minimal Pixels Regularization

Blind image deblurring is a long standing challenging problem in image processing and low-level vision. This work proposes an efficient and effective blind deblurring method, which utilizes a novel sparsity prior of local minimal pixels, namely patch-wise minimal pixels (PMP), to achieve accurate kernel estimation. In this paper, we will show that the PMP of clear images is much more sparse than that of blurred ones, and hence is very effective in discriminating between clear and blurred images. To efficiently exploit the sparsity of PMP in deblurring, an algorithm under the MAP framework to flexibly impose sparsity promotion on the PMP of the latent image is proposed. The sparsity promotion of PMP favors clear images over blurred ones in the deblurring process, and accordingly helps to yield more accurate kernel estimation. Extensive experiments demonstrate that the proposed algorithm can achieve state-of-theart performance on both natural and specific images. In terms of both deblurring quality and computational efficiency, the new algorithm is superior to other state-of-the-art methods. Code for reproducing the results of the new method reported in this work is available at https://github.com/FWen/deblur-pmp.git.