Virtual Fashion Try-On

Virtual try-on technology has become a promising option in the e-commerce industry.It helps users see how clothing would look on their own images before they buy.However, many existing models face issues with garment deformation, handling occlusion, and keeping garment textures accurate.This paper presents a complete virtual try-on framework that uses flow-infused attention for better garment alignment and a diffusion-based generator for high-quality image synthesis.We built the system with Python and React for the backend and frontend, respectively.Experiments on the VITON-HD dataset show improved structural similarity and visual realism compared to models like CP-VTON and VITON.