An Improved Criminisi Algorithm for Image Inpainting

In the image inpainting process, the repairing order and the matching accuracy of the Criminisi algorithm is not reasonable enough between target block and sample block. Aimed at these shortcomings, an improved algorithm based on the Criminisi algorithm is proposed in this paper. The Local Entropy of image is added to measure the size of the known information in target block when calculating the confidence term, and this paper redefined the data item taking into account the local variance and average gradient to measure the size of structural information which make the priority computation more reasonable. Meanwhile, structural similarity criterion is introduced to modify the matching strategy to search the best sample block. Lots of experimental results show that the proposed algorithm can enhance the continuity of the image structure and texture as well as minimize the artificial traces. So, it effectively improves the quality of the recovered image.