Next-Gen Virtual Try-Ons: U2NET-Driven Cloth Segmenting and Personalized Size Recommendation
Kunal Kankariya, Harashnabin Roy, B. Sowmiya · 2024
This project presents an end-to-end solution for virtual try-on applications, leveraging U2Net for clothing segmentation and image pattern modification. The U2Net model is fine-tuned on a labeled dataset for accurate clothing segmentation. Integrated into a virtual try-on system, the segmented clothing regions serve as a basis for applying diverse patterns and styles to enhance the user experience. The system incorporates image synthesis techniques for realistic pattern modification. With a user-friendly interface, it allows users to interact with and visualize different clothing styles in real-time. Rigorous testing and optimization ensure robust performance, culminating in a deployable solution for immersive and dynamic virtual try-on experiences.