Research map / 3D vision / Face and shape
EVA3D: Compositional 3D Human Generation from 2D Image Collections
Fangzhou Hong, Zhaoxi Chen, Yushi LAN, Liang Pan, Ziwei Liu
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 compositional body-part subnetworks (ablation) | EVA3D ablation baseline · DeepFashion | FID: 31.14 -> 17.81 | Improved |
| add compositional body-part subnetworks (ablation) | EVA3D ablation baseline · DeepFashion | Depth: 3.57 -> 5.02 | Hurt |
| add SMPL-template geometry and predict delta SDF (ablation) | EVA3D with compositional representation · DeepFashion | FID: 17.81 -> 15.62 | Improved |
| add SMPL-template geometry and predict delta SDF (ablation) | EVA3D with compositional representation · DeepFashion | Depth: 5.02 -> 3.69 | Improved |
| add pose-guided Gaussian sampling (ablation) | EVA3D with delta SDF · DeepFashion | Depth: 3.69 -> 3.04 | Improved |
| add pose-guided Gaussian sampling (ablation) | EVA3D with delta SDF · DeepFashion | FID: 15.62 -> 15.91 | Hurt |
| increase Gaussian pose-sampling σθ from 15° to 30° (ablation) | EVA3D with delta SDF · DeepFashion | Depth: 3.04 -> 2.58 | Improved |
| increase Gaussian pose-sampling σθ from 15° to 30° (ablation) | EVA3D with delta SDF · DeepFashion | FID: 15.91 -> 19.05 | Hurt |
| replace Gaussian pose sampling with uniform sampling (ablation) | EVA3D with delta SDF · DeepFashion | FID: 15.62 -> 25.82 | Hurt |
| replace original pose distribution with uniform sampling (ablation) | EVA3D with delta SDF · DeepFashion | Depth: 3.69 -> 2.92 | Improved |
| use pose-guided Gaussian sampling instead of the original dataset distribution (ablation) | EVA3D · DeepFashion | Depth: 3.69 -> 3.04 | Improved |
| use pose-guided Gaussian sampling instead of the original dataset distribution (ablation) | EVA3D · DeepFashion | FID: 15.62 -> 15.91 | Hurt |
Later papers that compared with this method
10 later accepted papers report EVA3D as a baseline in their results tables; 8 of them report a better number on most of the results they share with it.
- Generative Human Geometry Distribution (ICLR 2026 · better on 1 of 1 shared result)
- HumanLiff: Layer-wise 3D Human Diffusion Model (IJCV 2025 · better on 4 of 4 shared results)
- Disentangled Clothed Avatar Generation with Layered Representation (ICCV 2025 · better on 2 of 2 shared results)
- En3D: An Enhanced Generative Model for Sculpting 3D Humans from 2D Synthetic Data (CVPR 2024 · better on 4 of 4 shared results)
- AttriHuman-3D: Editable 3D Human Avatar Generation with Attribute Decomposition and Indexing (CVPR 2024 · better on 1 of 3 shared results)
- StructLDM: Structured Latent Diffusion for 3D Human Generation (ECCV 2024 · better on 2 of 3 shared results)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · better on 2 of 2 shared results)
- Gaussian Shell Maps for Efficient 3D Human Generation (CVPR 2024 · better on 0 of 1 shared result)
2 more not listed.
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- StyleSDF — StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generation (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
- En3D: An Enhanced Generative Model for Sculpting 3D Humans from 2D Synthetic Data (CVPR 2024 · similarity 0.90)
- Gaussian Shell Maps for Efficient 3D Human Generation (CVPR 2024 · similarity 0.88)
- HumanGaussian: Text-Driven 3D Human Generation with Gaussian Splatting (CVPR 2024 · similarity 0.87)
- PoseGen: Learning to Generate 3D Human Pose Dataset with NeRF (AAAI 2024 · similarity 0.87)
- SiTH: Single-view Textured Human Reconstruction with Image-Conditioned Diffusion (CVPR 2024 · similarity 0.87)
- GeneMAN: Generalizable Single-Image 3D Human Reconstruction from Multi-Source Human Data (NeurIPS 2025 · 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.
