Research on Character Image Inpainting based on Generative Adversarial Network

Image inpainting technology plays an important role in the process of digitalizing ancient literature. It helps to recover the partially missing or stained characters. Recently the Generative Adversarial Network (GAN) has shown remarkable success in the field of image inpainting. In this paper, we propose a new GAN model using the idea of edge recovery and optimize this model with spatial attenuation mask and conditional labelling to improve performance. Experiments show better results than the previous works in character image inpainting.