Research map / 3D vision / Face and shape
CanonicalFusion: Generating Drivable 3D Human Avatars from Multiple Images
Jisu Shin, Junmyeong Lee, Seongmin Lee, Min-Gyu Park, Jumi Kang, Ju Hong Yoon, Hae-Gon Jeon
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 |
|---|---|---|---|
| train on combined TH2.0 and RP data instead of a single dataset (ablation) | CanonicalFusion · RP | P2S (cm): 1.043 (TH2.0) / 0.901 (RP) -> 0.886 (TH2.0+RP) | Improved |
| use CanonicalFusion instead of SCANimate for canonicalization from posed scans (5 views) | CanonicalFusion · Canonical space SET1 | P2S (cm): 1.362 -> 0.244 | Improved |
| increase number of input views from 5 to 15 (ablation) | CanonicalFusion · Canonical space SET1 | P2S (cm): 0.244 -> 0.199 | Improved |
| use CanonicalFusion over 2K2K on monocular reconstruction | CanonicalFusion · RP | P2S (cm): 1.097 -> 0.886 | Improved |
| add pose refinement in forward skinning-based differentiable rendering (ablation) | CanonicalFusion · noisy-pose qualitative case | qualitative reconstruction | Improved |
| train on RP only vs TH2.0 only and test on TH3.0 (dataset mismatch) (ablation) | CanonicalFusion · TH3.0 | P2S (cm): 1.086 (RP) vs 1.074 (TH2.0) vs 1.072 (TH2.0+RP) | No clear effect |
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.
- SCANimate — SCANimate: Weakly Supervised Learning of Skinned Clothed Avatar Networks (CVPR 2021) · better on 5 of 5 shared results
- ICON — ICON: Implicit Clothed Humans Obtained From Normals (CVPR 2022) · better on 2 of 2 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
- FRESA: Feedforward Reconstruction of Personalized Skinned Avatars from Few Images (CVPR 2025 · similarity 0.89)
- 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting (CVPR 2024 · similarity 0.88)
- AniGS: Animatable Gaussian Avatar from a Single Image with Inconsistent Gaussian Reconstruction (CVPR 2025 · similarity 0.87)
- Arc2Avatar: Generating Expressive 3D Avatars from a Single Image via ID Guidance (CVPR 2025 · similarity 0.87)
- Human-3Diffusion: Realistic Avatar Creation via Explicit 3D Consistent Diffusion Models (NeurIPS 2024 · similarity 0.87)
- VINECS: Video-based Neural Character Skinning (CVPR 2024 · similarity 0.87)
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.
