Robust Multiframe Blind Image Deblurring

Multiframe deblurring estimates a sharp and clear image from a set of blurry and noisy observations. A variety of deblurring methods are proposed in the past two decades. However, most of them are sensitive to their assumed data and noise model, limiting their utility. This paper first reviews the related works and shows the non-robustness in the traditional model and then proposes a robust Bayesian multiframe blind deblurring model. Alternating minimization scheme is adopted to solve this model. Experimental results indicate the effectiveness of the proposed method.