Inpainting of multiple blind motion-blurred images based on multi-scale tight wavelet frame

Inpainting of degraded images due to motion blurring is one challenging problem in digital image processing. Many techniques on blind motion-blurred images have emerged in the literatures. One promising approach is to restore the clear image by using multiple images captured from the scene. However, the existing inpainting approaches via multiple images use the uniform model to restore the clear image, ignoring the fact that the uniform inpainting model to one type of blurred images does not imply that it is practical to the other type. In this paper, we propose different image inpainting models for different types of multiple blind motion-blurred images via multi-scale tight wavelet frame after roughly classifying the degraded images to texture and piecewise smooth images, and give the experimental results which show that the proposed method can not only improve the quality of different types of motion-blurred images, but remove the motion blurred image noise effectively.