This repository serves as a curated collection of papers, code, and resources for Virtual Try-On (VTON) models, focusing on image, video, and 3D-based approaches published after 2023. It aims to provide researchers and developers with an up-to-date overview of the latest advancements in AI-driven virtual fashion try-on technology.

How It Works

The collection highlights recent research trends, with a significant emphasis on diffusion models and transformer architectures. Papers are categorized by their primary modality (image, video, 3D) and include details on their publication venue and date. The focus on diffusion models suggests a trend towards generative approaches that offer high fidelity and controllability in simulating garment appearance on users.

Quick Start & Requirements

This repository is a curated list of research papers and associated code. To utilize the models, users would need to refer to the individual linked papers and their respective code repositories for installation and execution instructions. Prerequisites will vary per model but generally include Python environments, deep learning frameworks (PyTorch/TensorFlow), and potentially specific hardware like GPUs with CUDA support.

Highlighted Details

• Extensive coverage of diffusion model-based VTON techniques, including advancements in latent diffusion, attention mechanisms, and texture preservation.
• Inclusion of recent research from top-tier conferences like CVPR, ECCV, AAAI, and SIGGRAPH Asia.
• Categorization across image, video, and 3D try-on modalities, reflecting the breadth of current research.
• Features papers exploring controllability, zero-shot capabilities, and in-the-wild try-on scenarios.

Maintenance & Community

The project is ongoing and welcomes community contributions to expand and improve the collection. Users are encouraged to star and watch the repository for updates.

Licensing & Compatibility

The repository itself is a collection of links and does not have a specific license. The licensing and compatibility of individual models must be checked within their respective linked code repositories.

Limitations & Caveats

This is a curated list of research papers and not a single, runnable software package. Users must individually assess and implement each model, which may involve complex dependencies and setup procedures. The rapid pace of research means the collection is constantly evolving.