A unified framework for geometry and exemplar based image inpainting

Image inpainting is a process of retrieving missing portions of an image without introducing undesirable artifacts. Methods based on partial differential equations, variational formulations and diffusion are local in nature. They work well in propagating the geometry of images into target region. However, they fail to predict the values for large missing portions of the image. Exemplar methods try to solve the shortcomings of the earlier methods by generating texture as well. They use a confidence map to determine the order fill-in. Nonetheless, they often fail to capture the coherence with respect to the surrounding region. In recent times, the trend is to unify these techniques to get best of the both worlds. In this direction, we have proposed a unified framework of geometry and exemplar based image inpainting. A non-local variational image inpainting formulation is introduced using graph based regularization. The direction of inpainting is influenced by a confidence map. Gabor filter responses are utilized to automatically decide if diffusion or exemplar method to be used in the course of inpainting. Results show that our method performs better than the state-of-the art methods.