Image completion by a fast and adaptive exemplar-based image inpainting

A fast and adaptive method is proposed for completing missing parts caused by the removal of foreground or background elements from an image of natural scenery. Unlike most previous texture-synthesis based approaches using extensive search to find the appropriate texture, we synthesize the missing parts by image patches drawn from horizontally located areas because of the strong horizontal orientation in natural scenes. On the other hand, we present an adaptive scheme to determine the size of the template window for capturing features of various scales. A number of examples are given to demonstrate the effectiveness of our algorithm. Our results compare favorably to those obtained by existing techniques.