Exemplar Based Image Inpainting with Improved Priority Function

Digital image inpainting is a technique used for filling in the disappeared or spoiled regions of an image using information from the surrounding area. Image inpainting provides a means to restore damaged region of an image, such that the image looks complete and natural after the inpainting process. Exemplar-based algorithms are a popular technique for image inpainting, in which two important parts: deciding the fillingin order and selecting good exemplars. Criminisi’s approach is to search suitable patches from source regions to fill in the missing parts, but face a problem that: unsuitable selection of exemplars. To improve the problem, we introduce a self-governing approach through investigating the process of patches propagation. Defining a new separated priority function to propagate geometry and then synthesize image textures, aiming to well recover image geometry and textures. Compared to other approach, the improved priority definition can recover image geometry and textures well.