Try-On-Adapter: A Simple and Flexible Try-On Paradigm

TL;DR

This paper introduces Try-On-Adapter, a novel outpainting approach for virtual try-on that preserves key features and offers flexible control, outperforming previous inpainting-based methods in realism and quality.

Contribution

We propose Try-On-Adapter, an outpainting paradigm that improves virtual try-on by allowing flexible control and better preservation of details, surpassing existing inpainting methods.

Findings

Achieves state-of-the-art FID scores of 5.56 and 7.23 on VITON-HD.

Performs well even with low-quality face and garment images.

Outperforms previous inpainting-based methods in qualitative evaluations.

Abstract

Image-based virtual try-on, widely used in online shopping, aims to generate images of a naturally dressed person conditioned on certain garments, providing significant research and commercial potential. A key challenge of try-on is to generate realistic images of the model wearing the garments while preserving the details of the garments. Previous methods focus on masking certain parts of the original model's standing image, and then inpainting on masked areas to generate realistic images of the model wearing corresponding reference garments, which treat the try-on task as an inpainting task. However, such implements require the user to provide a complete, high-quality standing image, which is user-unfriendly in practical applications. In this paper, we propose Try-On-Adapter (TOA), an outpainting paradigm that differs from the existing inpainting paradigm. Our TOA can preserve the given face and garment, naturally imagine the rest parts of the image, and provide flexible control ability with various conditions, e.g., garment properties and human pose. In the experiments, TOA shows excellent performance on the virtual try-on task even given relatively low-quality face and garment images in qualitative comparisons. Additionally, TOA achieves the state-of-the-art performance of FID scores 5.56 and 7.23 for paired and unpaired on the VITON-HD dataset in quantitative comparisons.