RubbingGAN

RubbingGAN - Pytorch Implementation (AI for Content Creation Workshop CVPR 2022)

Introduction

This is the implementation of the paper "End-to-End Rubbing Restoration Using Generative Adversarial Network. Our goal is to restore the imcomplete calligraphy characters on the traidional Chinese stone inscriptions (Rubbings).

Dataset

We collect the dataset of ZhangMenglongBei, which used for rubbing restoration.

We have the following datasets:

• Training Dataset

• Testing Dataset

• Restoration (Incomplete) Dataset

Train

python RubbingGAN.py --dataroot PATH TO Training Dataset --valDataroot PATH TO Validation Dataset --exp PATH TO store outputs of generating images --log PATH TO store logs

Restoration Result

Reference

• pix2pix-pytorch https://github.com/taey16/pix2pixBEGAN.pytorch

• began—pytorch https://github.com/taey16/pix2pixBEGAN.pytorch