Research map / Image generation / Editing and style transfer
PG-VTON: Single-Pass Training-Free Virtual Try-On via Patch-Guided Reference Alignment
Guohao Zhao, Yuxin Peng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Patch-Anchored Identity Priming (PIP) (ablation) | ICL-only FLUX.1-fill · VITON-HD | FIDu: 13.393 -> 10.829 | Improved |
| add Reference-Aware Attention (RAA) (ablation) | ICL-only FLUX.1-fill · VITON-HD | FIDu: 13.393 -> 11.543 | Improved |
| add RAA on top of PIP (ablation) | PIP-only · VITON-HD | FIDu: 10.829 -> 9.873 | Improved |
| add RAA on top of PIP (ablation) | PIP-only · VITON-HD | LPIPSp: 0.100 -> 0.086 | Improved |
| use PG-VTON vs training-free OmniVTON | PG-VTON · VITON-HD | FIDu: 13.167 -> 9.873 | Improved |
| use PG-VTON vs trained Any2AnyTryon | PG-VTON · DressCode | FIDp: 5.478 -> 4.195 | Improved |
| PG-VTON vs Any2AnyTryon on LPIPS | PG-VTON · VITON-HD | LPIPSp: 0.085 vs 0.086 | No clear effect |
| PG-VTON vs OmniVTON on StreetTryOn FID | PG-VTON · StreetTryOn | FID: St2St 23.470 -> 21.028; M2S 33.450 -> 37.098 | Mixed |
| PG-VTON vs OmniVTON on CMMD | PG-VTON · StreetTryOn S2S | CMMD: 0.599 -> 0.263 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Any2AnyTryon — Any2AnyTryon: Leveraging Adaptive Position Embeddings for Versatile Virtual Clothing Tasks (ICCV 2025) · better on 7 of 8 shared results
- OmniVTON — OmniVTON: Training-Free Universal Virtual Try-On (ICCV 2025) · better on 8 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- BooW-VTON: Boosting In-the-Wild Virtual Try-On via Mask-Free Pseudo Data Training (CVPR 2025 · similarity 0.91)
- Incorporating Visual Correspondence into Diffusion Model for Virtual Try-On (ICLR 2025 · similarity 0.91)
- MV-VTON: Multi-View Virtual Try-On with Diffusion Models (AAAI 2025 · similarity 0.91)
- M&M VTO: Multi-Garment Virtual Try-On and Editing (CVPR 2024 · similarity 0.91)
- Correpondence Alignment For Improved Virtual Try-On (NeurIPS 2026 · similarity 0.89)
- Enhancing Virtual Try-On with Synthetic Pairs and Error-Aware Noise Scheduling (CVPR 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
