Dynamic-Clustering Extreme Intensity Prior Based Blind Image Deblurring

Xiaopan Li, Shiqian Wu, Shoulie Xie, Sos S. Agaian

Abstract: In blind image deblurring, feasible solutions have been obtained by exploiting image prior information such as dark channel prior, extreme channel prior, and local minimal intensity prior. The performance highly depends on these priors, which may have poor adaptability to different image contents in real-world applications. For example, these priors only consider the changes in local minimal and maximal intensity pixels in the blurring process and ignore the difference between these changes. In this paper, we propose a novel blind image deblurring approach based on dynamic-clustering extreme intensity prior. Specifically, the patch-wise maximal pixels (PMaxP) prior and patch-wise minimal pixels (PMinP) prior are employed and clustered into two by applying fuzzy c-means (FCM) clustering. The regularizations impose the sparsity inducing on either PMaxP prior or PMinP prior in each patch. Extensive experimental results demonstrate that the proposed method outperforms the state-of-the-art image deblurring algorithms on synthetic and real-world images.