Controllable flow fields for high-fidelity virtual try-on with diffusion models
Multimedia Systems · 2026-10 · 32 (9)
DOI: 10.1007/s00530-026-02660-9 ↗Kurz gefasst
Fachartikel in der Zeitschrift Multimedia Systems, online erschienen am 21.09.2026. Verfasst von Kun Li und 8 weiteren Autor:innen. Die Arbeit nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: City University of Hong Kong.
Abstract (Originalsprache)
Virtual try-on is a challenging task that seeks to create realistic images of individuals wearing specific garments and has significant applications in e-commerce and virtual fashion. Existing approaches have made considerable advancements but come with large memory requirements and fail to preserve subtle garment details. To improve on these limitations, we present a new end-to-end diffusion-based framework that allows for more efficient and controllable detail generation. The key component of our approach is the COFFLOW module that extracts precise spatial correspondence from the target garment to the person image and incorporate this into the diffusion process as guidance for image synthesis. We also introduce a blending module that facilitates better visual quality in unmasked areas like face and limbs. We train our model freezing the pretrained diffusion U-Net and updating only the COFFLOW module, leading to a significantly lower cost during training. We evaluate our method on two standard benchmarks, VITON-HD and DressCode. Quantitatively and qualitatively, we show that our method significantly improves upon the prior diffusion-based methods, with better garment fitting, texture preservation, and realistic pose fitting, all while requiring lower computational resources.
Förderung
- City University of Hong Kong
Zitieren
APA
Li, K., Po, L., Yu, W., Xue, Y., Wu, H., Liu, Y., Jiang, Z., Xu, X., & Wang, Y. (2026). Controllable flow fields for high-fidelity virtual try-on with diffusion models. Multimedia Systems, 32(9). https://doi.org/10.1007/s00530-026-02660-9
Vancouver
Li K, Po L, Yu W, Xue Y, Wu H, Liu Y, et al. Controllable flow fields for high-fidelity virtual try-on with diffusion models. Multimedia Systems. 2026;32(9). doi:10.1007/s00530-026-02660-9
BibTeX
@article{Li2026,
author = {Li, Kun and Po, Lai-Man and Yu, Wing-Yin and Xue, Yu and Wu, Haoxuan and Liu, Yuyang and Jiang, Zeyu and Xu, Xuyuan and Wang, Yexin},
title = {Controllable flow fields for high-fidelity virtual try-on with diffusion models},
journal = {Multimedia Systems},
year = {2026},
volume = {32},
number = {9},
doi = {10.1007/s00530-026-02660-9}
}
EUVIMED