Yunnan Tibetan Buddhist Mural Restoration Based on Entropy Reduction Strategy and Boundary Structure Factor

Aiming at the problems of repair redundancy and structure breakage in the current image inpainting algorithm based on sample block in the inpainting results of Tibetan Buddhist mural images in Yunnan, an improved algorithm based on the strategy of entropy reduction and boundary structure factor is proposed. Firstly, the selection rule of sample block is changed according to the entropy reduction strategy to improve the overall proportion of known information in the inpainting process. Then the boundary structure factor is introduced into the priority, and the repair order is adjusted according to the curvature information. The update rule of confidence value is defined by comprehensively considering the distance between the repaired pixel and the source region and the impact of the overall repair order. Experimental results show that the proposed algorithm can effectively improve the phenomenon of irrational filling and obtain high-quality restoration of mural images.