Grid Warping Postprocessing for Linear Motion Blur in Images

Abstract

The paper presents a method for linear motion blur and out-of-focus blur suppression in photographic images. Conventional image deconvolution algorithms usually have a regularization parameter that specifies a trade-off between incomplete blur removal and high probability of artifacts like ringing and noise. The idea of the proposed image deblurring method is to apply grid warping approach to improve image sharpness after conventional image deconvolution algorithms used with strong regularization. Grid warping algorithm moves pixels in edge neighborhood area towards the edges making them sharper without introducing artifacts like ringing and noise. The proposed method is expected to have the same sharpness but decreased risk of artifacts compared to standalone image deconvolution. In order to validate the proposed scheme, we have applied it to artificially blurred images and images with real blur, with different levels of noise and blur radii, directions and lengths.