BUILDING A FASHION E-COMMERCE WEBSITE INTEGRATED WITH A VIRTUAL FITTING ROOM

Abstract

The rapid development of e-commerce has transformed the fashion industry, yet online shopping continues to face significant limitations—most notably the inability for customers to visualize how garments would appear on their own bodies. This thesis presents a fashion e-commerce platform integrated with an AI-powered virtual try-on feature designed to address this challenge and enhance user confidence in purchasing decisions. The system provides essential shopping functionalities, including product browsing, filtering, reviewing, cart and order management, along with an administrative dashboard for monitoring revenue, managing products, handling inventory, promotional campaigns, and customer activity. A central contribution of the study is the implementation of a virtual try-on mechanism using Google Gemini’s generative AI and Veo 3.1 video preview capabilities. By enabling users to upload personal images and generating realistic outfit simulations, the system delivers a more intuitive and immersive shopping experience. The platform is developed using a decoupled architecture comprising a NodeJS/Express backend with MySQL and a ReactJS/TailwindCSS frontend. Experimental evaluation indicates that the AI-generated try-on results achieve satisfactory realism and processing performance, increasing user engagement and reducing uncertainty. The findings demonstrate that integrating Generative AI into fashion e-commerce systems not only improves customer experience but also provides valuable tools for business optimization.