In this paper, we address the problem of estimating and removing non-uniformmotion blur from a single blurry image. We propose a deep learning approach topredicting the probabilistic distribution of motion blur at the patch levelusing a convolutional neural network (CNN). We further extend the candidate setof motion kernels predicted by the CNN using carefully designed imagerotations. A Markov random field model is then used to infer a densenon-uniform motion blur field enforcing motion smoothness. Finally, motion bluris removed by a non-uniform deblurring model using patch-level image prior.Experimental evaluations show that our approach can effectively estimate andremove complex non-uniform motion blur that is not handled well by previousapproaches.
展开▼
