An Improved Blind Deconvolution Algorithm of Motion Blurred Image

Abstract:

An improved blind deconvolution algorithm has been proposed to tackle the image blurring caused by movement. Firstly, take the model to analyze the blurred image and identify the high-frequency of the blurred image through the reasonable pre-estimation of blind kernel function. Then, further estimate the blurring matrix K by the use of the high-frequency while the high-frequency of the blurred image and the accuracy of the blurred image matrix have been improved gradually through normalized iteration. Finally, the clear image has been gained by the use of blind matrix deconvolution. Experimental results show that by the use of this algorithm the image has better subjective visual effect. The proposed algorithm has better restoration effect on the fused motion blurred image and the natural camera motion blurred image.

• Periodical: Applied Mechanics and Materials (Volumes 668-669)
• Pages: 1045-1049
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.668-669.1045
• Online since: October 2014
• Authors: Lei Yang, Ze Feng Wang, Hui Nan Guo, Guang Sen Liu, Hao Wang, Li Nao Tang, Hong Tao Yang
• Keywords: Blind Deconvolution, Blind Kernel, Image Restoration, Motion Blurred Image

References

[1] Liao Yong-zhong, Cai Zi-ming, He Xiang-hua. Fast algorithm for motion blurred image blind deconvolution[J]. Optics and Precision Engineering, 2013, 21(10): 2688- 2695.

DOI: 10.3788/ope.20132110.2688

[2] MISKIN J. Ensemble Learning for Independent Component Analysis[D]. Cambridge: University of Cambridge. (2000).

[3] Dilip Krishnan, Terence Tay, Rob Fergus. Blind Deconvolution Using a Normalized Sparsity Measure[C]. CVPR 2011: 233-240.

DOI: 10.1109/cvpr.2011.5995521

[4] R. Fergus, B. Singh, A. Hertzmann, et al. Removing camera shake from a single photograph[C]. SIGGRAPH, 25: 787–794, (2006).

DOI: 10.1145/1179352.1141956

[5] W. Yin, S. Osher, D. Goldfarb, et al. Bregman iterative algorithms for l1-minimization with applications to compressed sensing[C]. SIAM J. on Img. Sciences, 1(1): 143–168, (2008).

DOI: 10.1137/070703983

[6] A. Beck and M. Teboulle. A fast iterative shrinkage-thresholding algorithm for linear inverse problems[C]. SIAM J. on Img. Sciences, 183–202, (2009).

DOI: 10.1137/080716542

[7] A. Levin, R. Fergus, F. Durand, et al. Image and depth from a conventional camera with a coded aperture[C]. SIGGRAPH, 26(3): 70, (2007).

DOI: 10.1145/1275808.1276464

[8] O. Whyte, J. Sivic, A. Zisserman, et al. Non-uniform deblurring for shaken images[C]. CVPR, (2010).

DOI: 10.1109/cvpr.2010.5540175

[9] KRISHNAN D, FERGUS R. Fast image deconvolution using hyper-Laplacian priors[J]. Advances in Neural Information Processing Systems, 2009, 22: 1-9.

[10] Wang Guo-dong, Xu Jie, Pan Zhen-kuan, et al. Blind image restoration based on normalized hyper laplacian prior term[J]. Optics and Precision Engineering, 2013, 21(5): 1340-1348.

DOI: 10.3788/ope.20132105.1340