Motion deblurring from a single image using gradient enhancement

Authors:
Jiahua Chen;Zhifeng Xie;Bin Sheng;Lizhuang Ma
Affiliations:
Shanghai Jiaotong University, Shanghai, P. R. China;Shanghai Jiaotong University, Shanghai, P. R. China;Shanghai Jiaotong University, Shanghai, P. R. China;Shanghai Jiaotong University, Shanghai, P. R. China
Venue:
Proceedings of the 10th International Conference on Virtual Reality Continuum and Its Applications in Industry
Year:
2011

Citing 10
Cited 0

Extrapolation, Interpolation, and Smoothing of Stationary Time Series

Extrapolation, Interpolation, and Smoothing of Stationary Time Series
Removing camera shake from a single photograph

ACM SIGGRAPH 2006 Papers
Image and depth from a conventional camera with a coded aperture

ACM SIGGRAPH 2007 papers
Total variation minimizing blind deconvolution with shock filter reference

Image and Vision Computing
A Closed-Form Solution to Natural Image Matting

IEEE Transactions on Pattern Analysis and Machine Intelligence
High-quality motion deblurring from a single image

ACM SIGGRAPH 2008 papers
Fast motion deblurring

ACM SIGGRAPH Asia 2009 papers
Correction of Spatially Varying Image and Video Motion Blur Using a Hybrid Camera

IEEE Transactions on Pattern Analysis and Machine Intelligence
Two-phase kernel estimation for robust motion deblurring

ECCV'10 Proceedings of the 11th European conference on Computer vision: Part I
Single Image Haze Removal Using Dark Channel Prior

IEEE Transactions on Pattern Analysis and Machine Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

Motion deblurring is one of the recovery problems in image restoration, which remains several challenges in kernel estimation and blind deconvolution. This paper proposes a new optimization method for estimating the blurring kernel by gradient enhancement, which can iteratively solve a uniform deblur model. In this model, the point-spread-function(PSF) can be accurately estimated and refined by gradually enhancing the image gradients. Our approach includes following steps: edge-preserving gradient enhancement, edge selection, kernel estimation and refinement, fast non-blind deconvolution. The edge-preserving gradient enhancement can restore sharp edges while have no effect in flat regions. Combined with the edge selection, it greatly helps to estimate the kernel. To improve its speed performance, the estimation and deconvolution steps are executed in frequency domain. Experimental results demonstrate that our method can efficiently produce an accurate blur kernel and a restored image with fine image details.