Image deblurring via enhanced low-rank prior

Wenqi Ren, Xiaochun Cao, Jinshan Pan, Xiaojie Guo, Wangmeng Zuo, and Ming-Hsuan Yang

Abstract

Low rank matrix approximation has been successfully applied to numerous vision problems in recent years. In this paper, we propose a novel low rank prior for blind image deblurring. Our key observation is that directly applying a simple low rank model to a blurry input image significantly reduces blur even without using any kernel information, while preserving important edge information. The same model can be used to reduce blur in the gradient map of a blurry input. Based on these properties, we introduce an enhanced prior for image deblurring by combining the low rank prior of similar patches from both the blurry image and its gradient map. We employ a weighted nuclear norm minimization method to further enhance the effectiveness of low rank prior for image deblurring, which retains the dominant edges and eliminates fine texture and slight edges in the intermediate images for kernel estimation. In addition, we evaluate the proposed enhanced low rank prior for both uniform and non-uniform deblurring. Quantitative and qualitative experimental evaluations demonstrate that the proposed algorithm performs favorably against the state-of-theart deblurring methods.

(a) Our observation

(b) Main step

Paper

[Paper]

Code

[Matlab Code]

Citation

@article{Ren-TIP-2016,

author = {Ren, Wenqi and Cao, Xiaochun and Pan, Jinshan and Guo, Xiaojie and Zuo, Wangmeng and Yang, Ming-Hsuan},

title = {Image Deblurring via Enhanced Low Rank Prior},

journal = {IEEE Transcations on image processing},

volume = {25},

number = {7},

pages = {3426--3437},

yers = {2016}

}