Research map / 3D vision / Face and shape
Zero-Shot Reconstruction of Animatable 3D Avatars with Cloth Dynamics from a Single Image
Joohyun Kwon, Geonhee Sim, Gyeongsik Moon
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 Dynamic Transformer (ablation) | DynaAvatar · 4D-Dress (ablation) | PSNR: 22.57 -> 23.62 | Improved |
| add Dynamic Transformer (ablation) | DynaAvatar · 4D-Dress (ablation) | SSIM: 0.952 -> 0.958 | Improved |
| add Dynamic Transformer (ablation) | DynaAvatar · 4D-Dress (ablation) | LPIPS: 0.068 -> 0.062 | Improved |
| add static-to-dynamic knowledge transfer with LoRA (vs training Static Transformer from scratch) (ablation) | DynaAvatar · DNA-Rendering / 4D-Dress (qualitative) | texture preservation | Improved |
| add LoRA (vs full fine-tuning of pretrained Static Transformer) (ablation) | DynaAvatar · DNA-Rendering / 4D-Dress (qualitative) | texture preservation | Improved |
| add DynaFlow loss (ablation) | DynaAvatar · DNA-Rendering / 4D-Dress (qualitative) | cloth motion and boundary quality | Improved |
| fully reannotate dataset with accurate SMPL-X fits (vs original annotations) (ablation) | DynaAvatar · 4D-Dress | PSNR: 20.72 -> 23.74 | Improved |
| reannotate only originally available frames (vs original annotations) (ablation) | DynaAvatar · 4D-Dress | PSNR: 20.72 -> 20.91 | No clear effect |
| reannotate all frames rather than only available frames (ablation) | DynaAvatar · 4D-Dress | PSNR: 20.91 -> 23.74 | Improved |
| use DynaAvatar instead of LHM | single-image avatar · DNA-Rendering | PSNR: 17.42 -> 19.45 | Improved |
| use DynaAvatar instead of IDOL | single-image avatar · Actors-HQ | PSNR: 20.93 -> 21.38 | Improved |
| use DynaAvatar instead of LHM | single-image avatar · DNA-Rendering | face consistency: 0.697 -> 0.712 | 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.
- PERSONA — PERSONA: Personalized Whole-Body 3D Avatar with Pose-Driven Deformations from a Single Image (ICCV 2025) · better on 9 of 9 shared results
- IDOL — IDOL: Inertial Deep Orientation-Estimation and Localization (AAAI 2021) · better on 9 of 9 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
- 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting (CVPR 2024 · similarity 0.90)
- MPMAvatar: Learning 3D Gaussian Avatars with Accurate and Robust Physics-Based Dynamics (NeurIPS 2025 · similarity 0.89)
- DLCA-Recon: Dynamic Loose Clothing Avatar Reconstruction from Monocular Videos (AAAI 2024 · similarity 0.89)
- GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians (CVPR 2024 · similarity 0.89)
- Vid2Avatar-Pro: Authentic Avatar from Videos in the Wild via Universal Prior (CVPR 2025 · similarity 0.89)
- FRESA: Feedforward Reconstruction of Personalized Skinned Avatars from Few Images (CVPR 2025 · 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.
