Research map / Video / Video generation
The devil is in the details: Enhancing Video Virtual Try-On via Keyframe-Driven Details Injection
Qingdong He, Xueqin Chen, Yanjie Pan, Peng Tang, Pengcheng Xu, Zhenye Gan, Chengjie Wang, Xiaobin Hu, Jiangning Zhang, Yabiao Wang
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 garment distillation component D (ablation) | KeyTailor · ViT-HD ablation | VFID_I (paired): 22.5241 -> 7.5267 | Improved |
| add garment details enhancement module (ablation) | KeyTailor · ViT-HD ablation | VFID_I (paired): 19.8968 -> 7.5267 | Improved |
| add collaborative background optimization (ablation) | KeyTailor · ViT-HD ablation | VFID_I (paired): 17.2134 -> 7.5267 | Improved |
| add instruction-guided keyframe sampling (ablation) | KeyTailor · ViT-HD ablation | VFID_I (paired): 16.2586 -> 7.5267 | Improved |
| add garment-area ratio score Sr(f) (ablation) | KeyTailor · ViT-HD ablation | VFID_I (paired): 17.2568 -> 7.5267 | Improved |
| use multiple keyframes instead of one (ablation) | KeyTailor · ViT-HD ablation | VFID_I (paired): 16.3856 -> 7.5267 | Improved |
| KeyTailor vs MagicTryOn | KeyTailor · ViT-HD test | VFID_I (paired): 14.0587 -> 7.5267 | Improved |
| fine-tune additionally on ViT-HD | KeyTailor · ViViD | VFID_I (paired): 9.5285 -> 8.2164 | Improved |
| fine-tune DreamVVT on own data vs KeyTailor | KeyTailor vs DreamVVT · ViViD | VFID_R (paired): 0.2549 vs 0.2761 | Mixed |
| use LoRA detail injection instead of added interaction modules | KeyTailor · Trainable params | params (B): 16.4460 -> 0.2057 | Improved |
| KeyTailor on image try-on | KeyTailor · VITON-HD | LPIPS: 0.0477 -> 0.0566 | Hurt |
Closest papers by meaning
- InstructVVT: Instruction-Driven Video Virtual Try-On without Auxiliary Spatial Priors (NeurIPS 2026 · similarity 0.92)
- Pursuing Temporal-Consistent Video Virtual Try-On via Dynamic Pose Interaction (CVPR 2025 · similarity 0.91)
- SwiftTry: Fast and Consistent Video Virtual Try-On with Diffusion Models (AAAI 2025 · similarity 0.91)
- FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization (NeurIPS 2026 · similarity 0.88)
- Cascaded Diffusion Models for Virtual Try-On: Improving Control and Resolution (AAAI 2025 · similarity 0.88)
- OpenVTON-Bench: A Large-Scale High-Resolution Benchmark for Controllable Virtual Try-On Evaluation (NeurIPS 2026 · similarity 0.88)
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.
