Fast Deblurring Method for Removing Uniform Motion Blur

Abstract

We propose a fast deblurring method to remove uniform motion blur. The proposed method is composed of a gradual latent estimation followed by an image refinement. The gradual latent estimation consists of a prediction phase and estimation phases. In the prediction phase, we use the guided image filtering to remove unusual edges. In the estimation phase, the latent and the point spread function are estimated by simplified estimation process. Then the estimated latent is refined to remove noises. We have evaluated the proposed method using public-domain data. It is shown that the proposed method is 6 times faster than conventional methods.