Exemplar-Based Image Inpainting Using Structure Consistent Patch Matching

摘要

Image inpainting reconstructs lost or deteriorated parts of images according to the information of surrounding regions. Criminisi has proposed an effective exemplar-based inpainting algorithm, which has the advantages of both texture synthesis and diffusion-based inpainting. Yet, it has its own flaws of fast priority dropping and visual inconsistency. In this paper, we propose a space varying updating strategy for the confidence term to improve the filling priority estimation and a structure consistent patch matching to take the difference distribution of source and target patches into account. Experimental results have demonstrated the improvement of our proposed method.