Noisy Radial Deblurring Using Natural Image Prior

Radial blurring, a challenging problem in image processing, is a special case of the space-variant motion blur. Previous methods mainly focus on space-invariant motion blurring, which could not be utilized directly in the space-variant motion deblurring. In this paper, we proposed a modified noisy radial deblurring method based on polar coordinate system, using the global and local image priors. The experiment results show that our proposed algorithm can denoise the additive noise and suppress the undesirable ring like ringing artifacts successfully, while the objective evaluations and subjective visual effects of the restoration images are improved obviously.