Neural Embedding of an Iterative Deconvolution Algorithm for Motion Blur Estimation and Removal

This paper introduces a new learning-based approach to motion blur removal. A local linear motion model is first estimated at each pixel using a convolutional neural network (CNN) in a regression setting. These estimates are then used to drive an algorithm that casts non-blind, non-uniform image deblurring as a least-squares problem regularized by natural image priors in the form of sparsity constraints. This problem is solved by combining the alternative direction method of multipliers with an iterative residual compensation algorithm, with a finite number of iterations embedded into a second CNN whose trainable parameters are deconvolution filters. The second network outputs the sharp image, and the two CNNs can be trained together in an end-to-end manner. Our experiments demonstrate that the proposed method is significantly faster than existing ones, and provides competitive results with the state of the art on several synthetic and real datasets.