End-to-End Rubbing Restoration Using Generative Adversarial Networks

TL;DR

This paper introduces RubbingGAN, a generative adversarial network designed for restoring incomplete rubbing characters, utilizing a new dataset from the Zhang Menglong Bei to effectively repair damaged cultural heritage artifacts.

Contribution

The paper presents the first GAN-based method for rubbing restoration and constructs the first dataset for this purpose, advancing digital preservation techniques.

Findings

RubbingGAN effectively restores incomplete rubbing characters

The model repairs both slight and severe damages

Restoration is fast and accurate

Abstract

Rubbing restorations are significant for preserving world cultural history. In this paper, we propose the RubbingGAN model for restoring incomplete rubbing characters. Specifically, we collect characters from the Zhang Menglong Bei and build up the first rubbing restoration dataset. We design the first generative adversarial network for rubbing restoration. Based on the dataset we collect, we apply the RubbingGAN to learn the Zhang Menglong Bei font style and restore the characters. The results of experiments show that RubbingGAN can repair both slightly and severely incomplete rubbing characters fast and effectively.