An Improved Image Inpainting Method Based on Patch Matching and Texture Synthesis
Shaofang Lu, ChangMing Han · 2019
Compared with the pixel based image restoration algorithm, the digital image restoration algorithm based on sample blocks is more effective for the restoration of texture and structure information in large area damaged images. The key technology is to determine the repair order. In this paper, an improved priority value calculation method is proposed. The experimental results show that the improved method can further improve the repair order of the region to be repaired and fill the missing part of the image with the matching block in the improved order on the basis of the original algorithm. The unreasonable extension phenomenon is reduced, and the texture and structure information are more accurately distinguished, which makes the inpainting effect more coherent in vision.