Removing an Object Using MRF Patch-Based Image Inpainting

A methodology for context-aware patch-based image inpainting is presented in this paper. Here textural descriptors are used, which controls and accelerates the search for nearest patches. The input image is inpainted by using a non- parametric patch sampling. Here, we have used the Markov random fields (MRF) for getting nearest patches on input image for better results. The computational complexity is reduced by the inpainting of the input image and it also allow to work with the dominant orientations of image structures. Image super-resolution is applied so as to recover the details of missing areas from the low-resolution inpainted image. Natural images and texture synthesis are used to demonstrate the effectiveness of the proposed method.