Image Deblurring via Enhanced Low-Rank Prior

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 the 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, by retaining the dominant edges and eliminating fine texture and slight edges in intermediate images, allowing for better kernel estimation. In addition, we evaluate the proposed enhanced low-rank prior for both the uniform and the non-uniform deblurring. Quantitative and qualitative experimental evaluations demonstrate that the proposed algorithm performs favorably against the state-of-the-art deblurring methods.

Image Deblurring via Alternating Direction Method of Multipliers

“…many common forms of image prior used in this setting have a major drawback in that the minimum of the resulting cost function does not correspond to the true sharp solution [5].many methods are used to get a better result. Here introduces an image regularization technique.…”

“…Filters can be used to generate a high frequency versions [5].the filters can be Δx= [1,1] and Δy= [1,1].the cost for spatially invariant blurring can be;…”

“…This subproblem can be non convex due to the presence of regularization term [5].the denominator can fix the sub problem then it becomes convex regularized problem. once the kernal is estimated then a vereity of deconvolution method can be used to recover the original image.…”

“…A novel prediction step is introduced and it works with image derivatives rather than pixel values. In the prediction step [5], it uses a simple image processing techniques to predict strong edges from an estimated latent image, by using kernel estimation.…”

“…This approach proposes a novel low rank prior for blind image deblurring [5]. That means information about the blur is not known.…”