Deep learning model to identify Vietnamese historical figures

Abstract

This study focuses on developing a deep learning-based model for recognizing historical figures from Vietnam's rich history, using both real and synthesized images. Given the limited availability of authentic historical images, we address this challenge by employing advanced image augmentation techniques such as image mixing, diffusion, and generative models to create diverse and realistic representations of these figures. A ResNet-50 model is then trained to accurately identify and classify these historical figures based on the enhanced dataset. The results show that combining AI with innovative image augmentation methods significantly improves the accuracy and robustness of historical figure recognition. This approach offers a more dynamic and accessible way to represent Vietnam's historical figures, with potential applications in education and digital heritage preservation.