Booth Id:
ROBO019
Category:
Robotics and Intelligent Machines
Year:
2023
Finalist Names:
Sun, Gongbo (School: Beijing National Day School)
Abstract:
Rubbings are important world cultural heritages. It provides crucial roles in humanity like arts, history, and politics. However, due to the poor portion and weathering process, most rubbings left on the world are incomplete and need to be restored, which is a time-consuming and laboring task. To solve this problem, we propose a novel AI-based restoration method called RubbingGAN. We also collect two data sets: one from Xuan Mita Rubbing, and another from Zhang Menglong Rubbing, where both have high artistic values. We evaluate our models based on the FID index and MSE index. The results show that our model can restore both the slightly and severely incomplete characters successfully.
Awards Won:
Fourth Award of $500