Zheng-Chong/Awesome-Try-On-Models

What it solves

Virtual try-on (VTON) lets people see how clothes, shoes, or accessories look on them without physically wearing the items. This is done with AI generative models that digitally simulate the outfit on a person's image or video. The challenge is making the result look realistic and preserving garment details, body shape, and motion. This repository collects the latest research and models in this field, focusing on works published after 2023, so researchers and developers can quickly see what new approaches exist.

How it works

The repository is a curated list, organized into three categories: image-based models, video-based models, and 3D-based models. Each entry links to the paper (arXiv or conference page), the project page or demo when available, and the code repository if open-source. The list is kept up to date with new papers, and each entry shows the publication venue (e.g., CVPR, ECCV, NeurIPS) and date. The README explains that the project is ongoing and welcomes contributions to expand and improve the collection. It also asks readers to star and watch the repo if they find it useful.

Who it's for

This is for AI researchers, computer vision engineers, and developers working on fashion technology, virtual try-on, or generative models for human images. It is also useful for anyone who wants to keep track of the state of the art in VTON, whether for academic research, product development, or simply to explore what is possible. The list is highly technical and links to papers and code, so it is aimed at people who are comfortable with reading research papers and understanding model architectures.

Highlights

• Focuses exclusively on VTON models published after 2023, giving a current snapshot of the field.
• Covers three modalities: image-based, video-based, and 3D-based try-on.
• Each entry includes the publication venue, date, and links to paper, code, and demos when available.
• Open to community contributions to keep the list growing and up-to-date.
• The repository itself is maintained by Zheng-Chong, who also has several VTON projects in the list (e.g., CatVTON, FastFit).