Edge-Based Image Completing Guided by Region Segmentation
Minqin Wang, Guoqiang Han, Yongqiu Tu · 2008
This paper proposes a novel algorithm which simultaneously inpaints structures and textures of damaged images. In past years, people copy the pixels, which come from the surrounding neighborhood or selected examplar, into damaged region along isolation direction. This might cause such result as blur edges, which looks "unreasonable" to human eyes. We perform segmentation based on PCNN which can divide the image into different regions based on texture features and achieve boundaries and regions at same time. After segmenting, we reconstruct those damaged boundaries, and then do inpainting in each region based on texture synthesis. The algorithm performs well on edge restoring. In additional, it shortens the performing time due to searching examplar in each region not in whole image space.