Examplar-Based Image Inpainting Under Smoothness Constraint

In the traditional examplar-based algorithm, the image inpainting effect is greatly degraded when filling strong structure image, which causes the disconnecting phenomenon. To solve this problem, this paper proposes an examplar-based inpainting algorithm under smoothness constraint. The filling order was optimized by establishing a union energy function, introducing a novel partial differential equation constraint and using new matching criterion. Both theoretical analysis and simulation results show that this method can overcome many shortcomings of traditional methods, such as structural error propagation and whole structure losing and blurring, and realize comparative perfect inpainting effect for images with obvious changed structure and large-scale damaged.