An Efficient Blind Image Deblurring Algorithm

This paper presents a novel algorithm which concerns with the fast implement of blind image deblurring with a well-reconstructed original image. Firstly, we model both the original image and the blur utilizing the harmonic model in the Sobolev image space, based on which, the prior distributions of them are obtained; Secondly, the Gamma distribution is used as the prior distributions of the unknown parameters to incorporate more prior knowledge for blind image deblurring; Finally, we estimate the original image, the blur and the unknown parameters simultaneously and iteratively by the evidence analysis method. The experimental results show the efficiency and the competitive performance compared of the proposed algorithm with existing blind image deblurring methods.