Digitalizing the Dress-up Experience: An Exploration of Virtual Try-On for Traditional Chinese Costume
Yifan Lu, Maochun Zhang, Wenxuan Huang, Shouqin Guan · 2023
With the recent rapid developments in artificial intelligence (AI) technology, virtual try-on technology empowered by generative approaches has reached a level of maturity that is ripe for commercialization. However, the application of virtual try-on technology to traditional cultural costumes has not received adequate attention from research institutions and enterprises. To address this research gap, we present a novel virtual try-on system for traditional Chinese costume that incorporates diverse AI technologies, including generative models, pose estimation, human parsing, and image processing algorithms, to provide an accurate and immersive virtual try-on experience. In particular, we leverage five distinct techniques Human Keypoint Detection, Human Parsing, Densepose Parts Segmentation, Cloth Mask Extraction, and Clothing-Agnostic Representation - to realize a web service that can effectively handle the specific challenges of traditional Chinese costume, such as the need to preserve cultural authenticity and ensure accurate fit and drape. Finally, we conduct various experiments to evaluate the system’s performance and have received positive feedback from users, indicating that it provides a realistic and engaging virtual try-on experience. Our work contributes to the growing body of research on virtual try-on technology and has the potential to impact the way traditional cultural costumes are designed, marketed, and worn.