A novel image inpainting framework based on multilevel image pyramids

Image inpainting is the art of manipulating an image so that it visually unrecognizable way. A considerable amount research has been done in this area over last few years. However, state techniques does suffer from computational complexities and plausible results. This paper proposes a multi-level pyramid-based algorithm. The algorithm starts with coarsest level pyramid overpainting information transferred to subsequent levels until bottom gets inpainted. search strategy used based on hashing coherent which makes fast accurate. Also, space constrained propagated thereby reducing complexity Compared other methods; proposed inpaints target region better plausibility human vision conformation. Experimental results show achieves as compared techniques.