Abstract:Ancient Chinese books are of great significance to historical research and cultural inheritance. Unfortunately, many of these books have been damaged and corroded in the process of long-term transmission. The restoration by digital preservation of ancient books is a new method of conservation. Traditional character restoration methods ensure the visual consistency of character images through character features and the pixels around the damaged area. However, reconstructing characters often causes errors, especially when there is large damage in critical locations. Inspired by human's imitation writing behavior, a two-branch structure character restoration network EA-GAN (Example Attention Generative Adversarial Network) is proposed, which is based on a generative adversarial network and fuses reference examples. By referring to the features of the example character, the damaged character can be restored accurately even when the damaged area is large. EA-GAN first uses two branches to extract the features of the damaged and example characters. Then, the damaged character is restored according to neighborhood information and features of the example character in different scales during the up-sampling stage. To solve problems when the example and damaged character features are not aligned and the convolution receptive field is too small, an Example Attention block is proposed to assist in restoration. Qualitative and quantitative analysis experiments are carried out on a self-built dataset MSACCSD and real scene pictures. Compared with current inpainting networks, EA-GAN can get the correct text structure through the guidance of the additional example in the Example Attention block. The peak signal-to-noise ratio (PSNR) and the structural similarity (SSIM) value increased by 9.82% and 1.82% respectively. The learned perceptual image patch similarity (LPIPS) value calculated by Visual Geometry Group (VGG) network and AlexNet decreased by 35.04% and 16.36% respectively. Our method obtained better results than the current inpainting methods. It also has a good restoration effect in the face of untrained characters, which is helpful for the digital preservation of ancient Chinese books.

STNet: Structure and Texture-Guided Network for Image Inpainting

UCTGAN: Diverse Image Inpainting Based on Unsupervised Cross-Space Translation

Image Inpainting Based on Interactive Separation Network and Progressive Reconstruction Algorithm

Image Inpainting by Bidirectional Information Flow on Texture and Structure

High-Resolution Image Inpainting Based On Multi-Scale Neural Network

Face Image Inpainting Based on Generative Adversarial Network

Art Image Inpainting with Style-guided Dual-branch Inpainting Network

Image Inpainting Based on Patch-GANs

An Ancient Chinese Painting Restoration Method Based on Improved Generative Adversarial Network

Two-stage structure aware image inpainting based on generative adversarial networks

An image inpainting method based on generative adversarial networks inversion and autoencoder

Sgrgan: sketch-guided restoration for traditional Chinese landscape paintings

Bi-encoder Network with Structure-texture Consistency for Image Inpainting.

Image Inpainting Using Two-Stage Loss Function and Global and Local Markovian Discriminators

Multi-Scale Generative Adversarial Networks Image Inpainting Algorithm*

Multistage Semantic‐aware Image Inpainting with Stacked Generator Networks

Image Inpainting Via Conditional Texture and Structure Dual Generation

EA-GAN: restoration of text in ancient Chinese books based on an example attention generative adversarial network

Research on Character Image Inpainting based on Generative Adversarial Network

Chinese Ancient Paintings Inpainting Based on Edge Guidance and Multi-Scale Residual Blocks

A Novel Generative Image Inpainting Model with Dense Gated Convolutional Network