StableVITON: Learning Semantic Correspondence with Latent Diffusion Model for Virtual Try-On

TL;DR

StableVITON leverages a pre-trained diffusion model with novel attention mechanisms to improve virtual try-on by accurately preserving clothing details and generating high-quality, realistic images.

Contribution

It introduces zero cross-attention blocks and a new attention total variation loss to effectively learn semantic correspondence within the diffusion model for virtual try-on.

Findings

Outperforms baseline methods in qualitative assessments.

Achieves higher quantitative accuracy in clothing detail preservation.

Produces sharper, more precise clothing representations.

Abstract

Given a clothing image and a person image, an image-based virtual try-on aims to generate a customized image that appears natural and accurately reflects the characteristics of the clothing image. In this work, we aim to expand the applicability of the pre-trained diffusion model so that it can be utilized independently for the virtual try-on task.The main challenge is to preserve the clothing details while effectively utilizing the robust generative capability of the pre-trained model. In order to tackle these issues, we propose StableVITON, learning the semantic correspondence between the clothing and the human body within the latent space of the pre-trained diffusion model in an end-to-end manner. Our proposed zero cross-attention blocks not only preserve the clothing details by learning the semantic correspondence but also generate high-fidelity images by utilizing the inherent knowledge of the pre-trained model in the warping process. Through our proposed novel attention total variation loss and applying augmentation, we achieve the sharp attention map, resulting in a more precise representation of clothing details. StableVITON outperforms the baselines in qualitative and quantitative evaluation, showing promising quality in arbitrary person images. Our code is available at https://github.com/rlawjdghek/StableVITON.