TAMF-VTON: Texture-Aware Mask-Free Virtual Try-On via High-Fidelity Image Synthesis

Gaofeng He; Huamin Wang; Jie Wang; Qian He; Xiaogang Jin

REVIEW 4 major objections 5 minor 13 references

This paper claims that mask-free, multi-garment virtual try-on can match or beat mask-based state-of-the-art methods while preserving fine textures, using a frozen diffusion backbone plus lightweight expert adapters and a synthetic training

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-02 00:59 UTC pith:HF24YLJ6

load-bearing objection A coherent mask-free VTON system that combines known components into a plausible pipeline, but the headline numbers rest on an unvalidated synthetic training base and no release. the 4 major comments →

arxiv 2607.14807 v1 pith:HF24YLJ6 submitted 2026-07-16 cs.CV

TAMF-VTON: Texture-Aware Mask-Free Virtual Try-On via High-Fidelity Image Synthesis

Jie Wang , Qian He , Gaofeng He , Xiaogang Jin , Huamin Wang This is my paper

classification cs.CV

keywords virtual try-onmask-freediffusion modelmixture-of-expertsfrequency-domain lossdata curationimage editingmulti-garment

verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper claims that virtual try-on can be done without any segmentation or inpainting masks at inference time, even when several garments are transferred at once. The proposed system, TAMF-VTON, treats try-on as a task-specialized image-editing problem on a frozen pretrained diffusion model, adding only lightweight expert adapters, a frequency-domain loss, and a data pipeline that synthesizes training pairs by simulating the reverse of try-on. If the claim holds, e-commerce try-on becomes simpler to deploy: no fragile mask prediction, support for arbitrary garment combinations, and better preservation of logos, weaves, and patterns than existing methods. The paper reports improved scores on standard benchmarks and shows results on complex real-world scenes, while noting that extreme lighting can shift garment color and that newly exposed skin may not perfectly match the person.

Core claim

The central discovery is that the three obstacles to practical mask-free try-on—mask dependence, texture loss, and limited multi-garment support—can be removed jointly within one diffusion-based editing framework. TAMF-VTON feeds the person image, one or more garment images, and a text instruction into a multimodal diffusion transformer as concatenated tokens. A token-wise mixture-of-experts adapter, with a few low-rank branches and top-2 routing, injects try-on-specific behavior while leaving the base model frozen. A frequency-domain loss, applied to the Fourier spectrum of the reconstructed image and weighted to emphasize low-noise steps, pushes the model to keep high-frequency garment det

What carries the argument

The central mechanism is the adaptive inpainting data pipeline that synthesizes aligned training pairs from ordinary try-on datasets. Since no dataset provides 'before-wear' images, the paper creates them: a detector locates the reference garment and its aspect ratio, pose keypoints map the body layout, and a linear regressor predicts the garment's hemline in the target image. That predicted region, combined with the original garment's mask and with hands and feet protected, becomes the inpainting region for a retrained inpainting model, yielding a reference-agnostic base image. The model then learns to map base image plus reference garment image(s) plus text instruction back to the original

Load-bearing premise

The load-bearing premise is that the lightweight linear regressor, fed by garment detection and pose keypoints, predicts where the reference garment will fall in the target image accurately enough that the inpainted 'before-wear' bases are semantically aligned with the final try-on; if that prediction is wrong for complex poses, occlusions, or cross-category swaps, the model trains on misaligned or artifact-laden images and the mask-free gains vanish.

What would settle it

Human-audit a random sample of synthesized training pairs (base image, reference garment, target image) for cases where the reference garment is much longer or shorter than the one worn. If the predicted hemline is systematically off—e.g., a long coat's hemline lands above the waist or a crop top's hemline extends past the hips—then the model is being trained on misaligned supervision, and the mask-free performance claim would be expected to fail exactly on cross-length transfers.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

Inference no longer requires a segmentation or inpainting mask: the same model accepts flat-lay product photos, one item or several, and a short text instruction, and outputs the synthesized person image.
Multi-garment outfits can be composed in a single pass rather than by sequentially editing one garment at a time, avoiding accumulated artifacts.
Fine garment details—logos, weaves, patterns—are better preserved, attributed to the frequency-domain supervision, especially for dense textures.
The frozen-backbone MoE design keeps the base model's general editing skills intact while adding try-on expertise cheaply.
With 4-bit quantization the model runs in under 15 seconds per image on a consumer GPU, making e-commerce deployment practical.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the training-pair synthesis is as faithful as claimed, the same inverse-try-on trick could be applied to other garment-centric editing tasks such as try-off, textile retexturing, or style transfer, since it converts unpaired data into supervised pairs without manual annotation.
The frequency-domain loss, applied globally rather than on masked regions, suggests a general recipe for diffusion-based editing: enforce spectral consistency on the whole reconstructed image with a timestep-dependent weight, which may transfer to non-fashion editing tasks where texture fidelity matters.
A testable extension: because the MoE router learns token-dependent expert activation implicitly, one could probe whether the experts specialize by spatial region or by garment category; if they do, the router could be exposed for user control, e.g., forcing a 'texture expert' for patterned garments.
The reported improvement in unpaired FID and KID implies better distributional realism; if true, mask-free training on synthetic pairs may narrow the gap between studio benchmarks and in-the-wild e-commerce photos, though human preference studies would be the natural next check.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Desk Editor's Note

A coherent mask-free VTON system that combines known components into a plausible pipeline, but the headline numbers rest on an unvalidated synthetic training base and no release.

read the letter

Punchline: The paper presents a coherent mask-free VTON pipeline that combines three known components—token-wise MoE adapters, a full-image frequency loss, and adaptive inpainting for pseudo-pair synthesis—into a system that, if the numbers hold, would be a useful engineering step. But the headline numbers rest on an unvalidated synthetic-training-data assumption, and no code or data is released, so the empirical claims are not independently checkable.

What's actually new: the adaptive inpainting strategy (AIS) is the most original piece. Using YOLOv10 detection, OpenPose keypoints, and a linear regressor to predict hemline position, then retraining FitDiT with those masks to generate reference-agnostic bases, is a plausible way to create aligned mask-free training triplets. The full-image frequency domain loss with a timestep-dependent weight is also a sensible modification of prior masked frequency losses. The equations in Sec. 4.2 are internally consistent; Eq. (6) is a correct inversion of the flow-matching velocity. The ablations show clear gains from MoE and the frequency term, and the authors are honest about remaining limitations (illumination, exposed skin).

Where the soft spots are: the entire mask-free capability at inference depends on the synthetic I_base images being faithful before-wear proxies. That premise is never validated. There is no quantitative evaluation of the AIS masks, no comparison of synthetic bases against real captured bases, and no evidence that the regressor handles cross-category, long-coat, or loose-dress cases. If FitDiT leaves artifacts or the hemline is mispredicted, the model is trained to edit synthetic noise, and the SOTA metrics in Table 1 could be an artifact of inverting the specific synthetic pipeline. The train/test overlap is also unclear: the paper says it trains on a 20K set constructed from VITON-HD and DressCode without stating whether the standard test splits were excluded. No code or data is released; no error bars are reported; multi-garment results are qualitative only; and the frequency-loss weight λ and cutoff radius r are not sensitivity-analyzed. These are addressable, not fatal.

Who it's for: researchers working on virtual try-on or image editing in the diffusion-model space. It deserves a serious referee—the system is plausible and the AIS idea is worth examining—but the review should be conditional on the authors releasing code/data, documenting the data splits, and providing evidence that the synthetic bases are faithful proxies for real before-wear images.

Recommendation: send to peer review with major-revision expectations, insisting on artifact release and a direct validation of the synthetic-to-real gap.

Referee Report

4 major / 5 minor

Summary. The paper proposes TAMF-VTON, a mask-free virtual try-on framework built on the Qwen-Edit diffusion transformer. It introduces three components: (1) a lightweight Mixture-of-Experts (MoE) LoRA adaptation with top-2 token routing, claimed to preserve the base model's general editing ability; (2) a frequency-domain supervision loss that penalizes high-frequency spectral differences between the decoded prediction and ground truth; and (3) an adaptive inpainting pipeline that synthesizes mask-free training tuples by replacing the target person's garment with a randomly sampled garment using a retrained FitDiT prior. The authors report state-of-the-art SSIM, LPIPS, FID, and KID on VITON-HD and DressCode, support for single- and multi-garment transfer, and inference under 15 seconds on an RTX 4090 with INT4 quantization.

Significance. If the results are valid, TAMF-VTON would be a practically valuable contribution: it removes the need for inference-time segmentation masks, supports multi-garment composition, and achieves efficient deployment on consumer hardware. The architectural ideas are coherent and the derivations in Eqs. (6)–(9) are internally consistent. The qualitative examples are appealing. However, the central quantitative claim rests on a synthetic training-data pipeline whose faithfulness to real 'before-wear' images is not validated. Because the model is trained and evaluated on the same benchmark datasets without documented split provenance, the reported improvements over baselines may be inflated. The paper also provides no code, no error bars, and limited reproducibility details, which weakens confidence in the headline SOTA numbers.

major comments (4)

[§4.3, §4.3.1] The entire mask-free training set is built from synthetic I_base images generated by the retrained FitDiT prior using the adaptive inpainting masks. The paper treats these synthetic bases as faithful proxies for true before-wear images, but provides no quantitative validation of this assumption. Supplementary Fig. 1 only shows qualitative comparisons; there is no metric for mask quality, no distributional comparison (e.g., FID/LPIPS) between synthetic I_base and real before-wear images, and no failure analysis for cross-category, long-coat, or occluded cases. If the hemline regressor or the inpainting prior produces misaligned or artifact-laden bases, the model is trained on corrupted supervision, and the Table 1 improvements could reflect inverting the synthetic pipeline rather than performing mask-free try-on on real photographs. Please provide a validation of the synthetic bases or an
[§5.1, Table 1] The paper trains on a 20K synthetic dataset 'constructed from VITON-HD and DressCode' and then evaluates on the test sets of those same benchmarks. It is not stated whether the synthetic training samples are derived from the train split or the test split, nor whether any person identities or garment instances overlap between training and evaluation. This is a load-bearing data-hygiene issue: if the training tuples are synthesized from the test set's own persons or garments, the paired SSIM/LPIPS numbers may be favorably biased by identity leakage. The split provenance must be documented, including the number of unique identities/garments in training versus test.
[Table 1, §5.3] Quantitative results are reported without error bars, number of seeds, or statistical significance tests. Many of the claimed improvements are small (e.g., VITON-HD SSIM 0.913 vs. 0.898 for FitDiT; KID 0.265 vs. 0.342). Without confidence intervals or significance testing, it is unclear whether these differences are meaningful, especially given the unvalidated synthetic training data. Additionally, some baseline rows have missing FID/KID values; the comparisons are therefore incomplete. Please report mean ± std over multiple runs (or the evaluation protocol) and complete the baseline results.
[§5.1, 'Nunchaku acceleration engine'] The inference-speed claim of under 15 seconds on an RTX 4090 with INT4 quantization is not accompanied by measurement details: no warmup, no batch size, no explicit image resolution, no version of the acceleration engine, and no distinction between first-call and steady-state latency. Since practical deployment is a stated motivation, this should be specified. This is a secondary but non-negligible issue for the paper's practical claims.

minor comments (5)

[§5.2.1] Typo: 'We presents both single-garment and multi-garment try-on comparisons' should be 'We present'.
[§4.2.1] The sentence 'Since the ground-truth velocity is v_t = epsilon - z_0' is a fragment; rephrase for clarity.
[Figure 2] The overview figure is dense. Consider labeling the MoE router and the flow of the frequency loss more explicitly so that the three contributions are immediately visible.
[References] The text refers to the 'Nunchaku acceleration engine [25]', but reference [25] is 'SVDQuant'. Please clarify the relation or correct the citation.
[Reproducibility] Only a commercial project page is provided, not a code repository. If code is not released, state this explicitly and describe how the benchmarks were evaluated in sufficient detail for reproduction.

Circularity Check

0 steps flagged

No significant circularity: all training losses and benchmark metrics are anchored to real ground-truth images; synthetic training bases are an empirical proxy, not a definitional identity.

full rationale

TAMF-VTON's derivation chain is supervised against external ground truth at every key stage. The flow-matching loss (Eq. 4) is defined w.r.t. the true velocity v_t = eps - z0; the spectral loss (Eq. 9) compares the decoded estimate x0_hat = D(z_t - t*v_hat) against the real target image x0; and the benchmark metrics in Table 1 are computed on VITON-HD and DressCode test images with real ground-truth try-on outputs. The adaptive-inpainting data curation (Sec. 4.3) uses a mask-based VTON prior (retrained FitDiT) to synthesize I_base training inputs, but the training target I_target is the original real photograph, so the model is not asked to reproduce its own synthetic artifact as if it were a prediction. The linear regressor for hemline placement is a data-construction heuristic, not a fitted parameter whose output is later reported as a benchmark prediction. No load-bearing step is justified solely by a self-citation: the reference list contains no author-overlapping 'uniqueness' or ansatz-citation chain, and the ablations (MoE vs LoRA, with/without L_f) change only the trained model while holding the external test protocol fixed. The unvalidated synthetic-to-real fidelity of FitDiT-generated I_base is an empirical validity/correctness risk, not circularity; if I_base is misaligned or artifact-laden, the training signal is degraded, but the derivation does not reduce Eq. X to Eq. Y by construction. Similarly, the limitation that AIS-vs-NAIS is only qualitatively validated (Supplementary Fig. 1) is a missing-support concern, not a circular step. Hence no significant circularity; score 0.

Axiom & Free-Parameter Ledger

5 free parameters · 5 axioms · 0 invented entities

The central method rests on three hand-tuned or learned-but-underspecified components: the spectral loss hyperparameters, the MoE topology, and the inpainting-prior-based data synthesis. No invented physical or architectural entities are postulated.

free parameters (5)

spectral loss weight λ = not reported
Eq. 10 combines L_FM + λL_f; λ is chosen by hand but its value and sensitivity are not given.
high-frequency cutoff r = (h+w)/32
Eq. 7 defines the radial high-frequency mask; the cutoff is a hand-set heuristic.
timestep weighting exponent = (1-t)^2
Eq. 8; hand-set schedule suppressing spectral loss at high-noise timesteps; no ablation on the exponent.
MoE topology (rank, experts, top-k) = rank 64, 4 experts, top-2
Section 5.1; selected by ablations in supplementary Table 2, not derived from first principles.
hemline regressor weights = not reported
Section 4.3.1; a linear regressor predicts hemline position from YOLOv10 detection and OpenPose keypoints; training data and weights are not described.

axioms (5)

standard math Flow matching training objective with standard velocity definition v_t = ε − z0
Background from [28]; used to derive the x0 estimate in Eq. 6.
domain assumption Qwen-Edit MMDiT plus VAE provide a capable base for multimodal image editing
Section 4; the whole pipeline is fine-tuned on top of this model without validating its limits.
ad hoc to paper Synthetic base images from the retrained FitDiT inpainting prior are valid 'before-wear' inputs
Section 4.3; the training triplets are generated by inpainting; no real before-wear data is used.
domain assumption Spectral supervision on the VAE-decoded x0 is meaningful despite latent-space information loss
Section 4.2.1; the VAE decoder may blur high-frequency detail, limiting the FFT-based signal.
ad hoc to paper The hemline regressor generalizes across categories and poses
Section 4.3.1; no validation of the regressor is provided.

pith-pipeline@v1.3.0-alltime-deepseek · 16594 in / 11585 out tokens · 91114 ms · 2026-08-02T00:59:24.608448+00:00 · methodology

0 comments

read the original abstract

Recent diffusion-based virtual try-on (VTON) methods remain limited by their reliance on segmentation masks, insufficient preservation of fine-grained textures, and limited support for arbitrary multi-garment compositions. Consequently, existing approaches still face significant challenges in real-world e-commerce deployment. We present TAMF-VTON, a texture-aware, mask-free framework that enables high-fidelity image synthesis under practical unconstrained conditions. Our method requires no human parsing or inpainting masks at inference time and supports diverse garment styles, categories, and quantities, enabling the simultaneous transfer of multiple items while preserving body structure and intricate texture details. This is achieved through a unified generative pipeline with three key components: (1) a lightweight Mixture-of-Experts (MoE) adaptation scheme that enables efficient fine-tuning without compromising the base model's general editing capabilities; (2) a frequency-domain supervision mechanism that explicitly optimizes high-frequency spectral consistency to preserve high-fidelity textures; and (3) a robust data curation pipeline employing an adaptive inpainting strategy to simulate the inverse VTON process for high-quality training pair generation. Extensive experiments demonstrate that our approach outperforms state-of-the-art methods in both quantitative metrics and perceptual quality. Optimized for efficiency, the model achieves inference in under 15 seconds per image on an NVIDIA RTX 4090 with INT4 quantization. By combining mask-free operation, flexible multi-garment composition, faithful texture preservation, and efficient inference on consumer hardware, TAMF-VTON demonstrates a commercially viable solution for scalable deployment in real-world digital fashion scenarios. The project is available at https://www.style3d.ai/ai-photoshoot/virtual-clothing-try-on.

Figures

Figures reproduced from arXiv: 2607.14807 by Gaofeng He, Huamin Wang, Jie Wang, Qian He, Xiaogang Jin.

**Figure 1.** Figure 1: Qualitative results of TAMF-VTON across diverse virtual try-on scenarios. The first row shows single-garment transfer from a flat-lay product image to a [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗

**Figure 2.** Figure 2: Overview of the TAMF-VTON architecture. Multimodal inputs are encoded as visual and textual tokens, concatenated, and processed by the MoE [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Figure 3: Visualization of garment representations in the pixel and frequency [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

**Figure 4.** Figure 4: Pipeline for synthesizing mask-free VTON training samples. We first [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗

**Figure 6.** Figure 6: Ablation study on MoE adaptation. Compared with the LoRA base [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗

**Figure 7.** Figure 7: Ablation study on frequency-domain supervision. Removing the [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗

**Figure 8.** Figure 8: Qualitative comparison on VITON-HD and DressCode. TAMF-VTON preserves garment shape, body structure, and fine-grained texture details more [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗

**Figure 9.** Figure 9: Robust single-garment virtual try-on under complex backgrounds and subject-to-subject transfer settings. Best viewed when zoomed in. [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗

**Figure 10.** Figure 10: Robust multi-garment virtual try-on under complex backgrounds and subject-to-subject transfer settings. Best viewed when zoomed in. [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗

**Figure 1.** Figure 1: Comparison between non-adaptive inpainting strategies (NAIS) and the proposed adaptive inpainting strategy (AIS). Existing pipelines construct [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗

**Figure 2.** Figure 2: Qualitative comparison on VITON-HD and DressCode. Single-garment try-on results for upper-body, lower-body, and dress categories, transferring [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗

**Figure 3.** Figure 3: Qualitative comparison on VITON-HD and DressCode. Multi-garment try-on results transferring multiple flat-lay garments onto studio-captured [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗

**Figure 4.** Figure 4: Qualitative comparisons in challenging real-world scenarios, including subject-to-subject transfer, subjects with challenging body-shape variations, and [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗

**Figure 5.** Figure 5: Additional qualitative results of TAMF-VTON under challenging real-world scenarios, with a particular focus on multi-garment and outfit try-on. The [PITH_FULL_IMAGE:figures/full_fig_p016_5.png] view at source ↗

**Figure 6.** Figure 6: Ablation study on the proposed frequency-domain supervision. [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗

discussion (0)

Reference graph

Works this paper leans on

13 extracted references · 3 linked inside Pith

[1]

Seunghwan Choi, Sunghyun Park, Minsoo Lee, and Jaegul Choo. 2021. VITON- HD: High-Resolution Virtual Try-On via Misalignment-Aware Normalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 14131–14140

2021
[2]

Yisol Choi, Sangkyung Kwak, Kyungmin Lee, Hyungwon Choi, and Jinwoo Shin
[3]

Zheng Chong, Xiao Dong, Haoxiang Li, Shiyue Zhang, Wenqing Zhang, Xu- jie Zhang, Hanqing Zhao, Dongmei Jiang, and Xiaodan Liang. 2025. Catvton: Concatenation is all you need for virtual try-on with diffusion models.The International Conference on Learning Representations(2025)

2025
[4]

Chenghu Du, Shengwu Xiong, and Yi Rong. 2025. All Parts Matter: A Unified Mask-Free Virtual Try-On Framework. InProceedings of the IEEE/CVF Interna- tional Conference on Computer Vision (ICCV). 19525–19534

2025
[5]

Boyuan Jiang, Xiaobin Hu, Donghao Luo, Qingdong He, Chengming Xu, Jinlong Peng, Jiangning Zhang, Chengjie Wang, Yunsheng Wu, and Yanwei Fu. 2024. Fitdit: Advancing the authentic garment details for high-fidelity virtual try-on. arXiv preprint arXiv:2411.10499(2024)

Pith/arXiv arXiv 2024
[6]

Black Forest Labs. 2025. FLUX.2: Frontier Visual Intelligence. https://bfl.ai/blog/ flux-2

2025
[7]

Davide Morelli, Matteo Fincato, Marcella Cornia, Federico Landi, Fabio Cesari, and Rita Cucchiara. 2022. Dress Code: High-Resolution Multi-Category Virtual Try-On. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. 2231–2235

2022
[8]

Gemini Team, Rohan Anil, and Sebastian Borgeaud. 2025. Gemini: A Family of Highly Capable Multimodal Models. arXiv:2312.11805 [cs.CL] https://arxiv.org/ abs/2312.11805

Pith/arXiv arXiv 2025
[9]

Zhenchen Wan, Yanwu Xu, Dongting Hu, Weilun Cheng, Tianxi Chen, Zhao- qing Wang, Feng Liu, Tongliang Liu, and Mingming Gong. 2025. MFT-VITON: High-Fidelity Virtual Try-On with Minimal Input via a Mask-Free Transformer- Diffusion Model. In2025 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). 2006–2015. doi:10.1109/ICCVW69036.2025.00210

arXiv 2025
[10]

Chenfei Wu, Jiahao Li, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kun Yan, Sheng- ming Yin, Shuai Bai, Xiao Xu, Yilei Chen, et al . 2025. Qwen-image technical report.arXiv preprint arXiv:2508.02324(2025)

Pith/arXiv arXiv 2025
[11]

Yuhao Xu, Tao Gu, Weifeng Chen, and Arlene Chen. 2025. Ootdiffusion: Outfitting fusion based latent diffusion for controllable virtual try-on. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 8996–9004

2025
[12]

Xuanpu Zhang, Dan Song, Pengxin Zhan, Tianyu Chang, Jianhao Zeng, Qing- guo Chen, Weihua Luo, and An-An Liu. 2025. BooW-VTON: Boosting In-the- Wild Virtual Try-On via Mask-Free Pseudo Data Training. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 26399–26408. doi:10.1109/CVPR52734.2025.02458

arXiv 2025
[2024]

In European Conference on Computer Vision

Improving diffusion models for authentic virtual try-on in the wild. In European Conference on Computer Vision. Springer, 206–235

[1] [1]

Seunghwan Choi, Sunghyun Park, Minsoo Lee, and Jaegul Choo. 2021. VITON- HD: High-Resolution Virtual Try-On via Misalignment-Aware Normalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 14131–14140

2021

[2] [2]

Yisol Choi, Sangkyung Kwak, Kyungmin Lee, Hyungwon Choi, and Jinwoo Shin

[3] [3]

Zheng Chong, Xiao Dong, Haoxiang Li, Shiyue Zhang, Wenqing Zhang, Xu- jie Zhang, Hanqing Zhao, Dongmei Jiang, and Xiaodan Liang. 2025. Catvton: Concatenation is all you need for virtual try-on with diffusion models.The International Conference on Learning Representations(2025)

2025

[4] [4]

Chenghu Du, Shengwu Xiong, and Yi Rong. 2025. All Parts Matter: A Unified Mask-Free Virtual Try-On Framework. InProceedings of the IEEE/CVF Interna- tional Conference on Computer Vision (ICCV). 19525–19534

2025

[5] [5]

Boyuan Jiang, Xiaobin Hu, Donghao Luo, Qingdong He, Chengming Xu, Jinlong Peng, Jiangning Zhang, Chengjie Wang, Yunsheng Wu, and Yanwei Fu. 2024. Fitdit: Advancing the authentic garment details for high-fidelity virtual try-on. arXiv preprint arXiv:2411.10499(2024)

Pith/arXiv arXiv 2024

[6] [6]

Black Forest Labs. 2025. FLUX.2: Frontier Visual Intelligence. https://bfl.ai/blog/ flux-2

2025

[7] [7]

Davide Morelli, Matteo Fincato, Marcella Cornia, Federico Landi, Fabio Cesari, and Rita Cucchiara. 2022. Dress Code: High-Resolution Multi-Category Virtual Try-On. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. 2231–2235

2022

[8] [8]

Gemini Team, Rohan Anil, and Sebastian Borgeaud. 2025. Gemini: A Family of Highly Capable Multimodal Models. arXiv:2312.11805 [cs.CL] https://arxiv.org/ abs/2312.11805

Pith/arXiv arXiv 2025

[9] [9]

Zhenchen Wan, Yanwu Xu, Dongting Hu, Weilun Cheng, Tianxi Chen, Zhao- qing Wang, Feng Liu, Tongliang Liu, and Mingming Gong. 2025. MFT-VITON: High-Fidelity Virtual Try-On with Minimal Input via a Mask-Free Transformer- Diffusion Model. In2025 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). 2006–2015. doi:10.1109/ICCVW69036.2025.00210

arXiv 2025

[10] [10]

Chenfei Wu, Jiahao Li, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kun Yan, Sheng- ming Yin, Shuai Bai, Xiao Xu, Yilei Chen, et al . 2025. Qwen-image technical report.arXiv preprint arXiv:2508.02324(2025)

Pith/arXiv arXiv 2025

[11] [11]

Yuhao Xu, Tao Gu, Weifeng Chen, and Arlene Chen. 2025. Ootdiffusion: Outfitting fusion based latent diffusion for controllable virtual try-on. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 8996–9004

2025

[12] [12]

Xuanpu Zhang, Dan Song, Pengxin Zhan, Tianyu Chang, Jianhao Zeng, Qing- guo Chen, Weihua Luo, and An-An Liu. 2025. BooW-VTON: Boosting In-the- Wild Virtual Try-On via Mask-Free Pseudo Data Training. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 26399–26408. doi:10.1109/CVPR52734.2025.02458

arXiv 2025

[13] [2024]

In European Conference on Computer Vision

Improving diffusion models for authentic virtual try-on in the wild. In European Conference on Computer Vision. Springer, 206–235