AI-Based Innovative Design Research on Female Figure Modeling with Tang Dynasty Costumes

Abstract

As a quintessential representative of traditional Chinese costume, Tang Dynasty attire integrates indigenous elements from the Han and Wei dynasties with exogenous cultural influences from the Western Regions and Persia, constituting a distinctive sartorial system of significant academic and industrial value. Nevertheless, conventional manual restoration and design methodologies exhibit pronounced limitations in efficiency, accuracy, and scalability, rendering them inadequate to address the exigent demands of contemporary digital preservation and creative design. In recent years, generative artificial intelligence technologies—particularly diffusion models, Neural Radiance Fields (NeRF), and large language models—have achieved substantial breakthroughs, furnishing novel technical pathways for the digital reconstruction and innovative design of cultural heritage. Grounded in these advancements, this study takes Tang Dynasty female costumes as a case study to explore and optimize AI-assisted generation methodologies for depicting female figures in Tang Dynasty attire.