A new algorithm for improving Deblurring effects and addressing spatially-variant blur problems for image motion deblurring
Jun Xie, Weiyao Lin, Hongxiang Li, Kai Guo, Bin Jin, Yihao Zhang, Donghua Liu · 2011
Motion deblurring techniques have played important roles in many image processing applications. In this paper, a new algorithm is proposed for motion deblurring from a single image. The proposed algorithm introduces an anisotropic Patiral Differential Equation (PDE) method for latent image prediction and employs an adaptive optimization model in the blind deconvolution iterative step. Furthermore, in order to solve the spatially-variant blur problems, we also propose a Saliency-based Deblurring (SD) approach by introducing saliency extraction together with a compensate method for image deblurring. Experimental results have demonstrated the effectiveness of the proposed algorithm.