Research map / 3D vision / Face and shape
Nonlinear 3D Face Morphable Model
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace linear PCA texture bases with a nonlinear texture decoder | 3DMM · In-the-wild face texture | L1 reconstruction error: 0.103 -> 0.066 | Helped |
| Replace linear PCA shape bases with a nonlinear shape decoder | 3DMM · 10 held-out 3D face scans | NME: 0.0241 -> 0.0196 | Helped |
| Use nonlinear instead of linear shape representation at latent size l_S=40 (ablation) | 3DMM · 10 held-out 3D face scans | NME: 0.0321 -> 0.0277 | Helped |
| Use nonlinear instead of linear shape representation at latent size l_S=80 (ablation) | 3DMM · 10 held-out 3D face scans | NME: 0.0279 -> 0.0236 | Helped |
| Use nonlinear instead of linear shape representation at latent size l_S=160 (ablation) | 3DMM · 10 held-out 3D face scans | NME: 0.0241 -> 0.0196 | Helped |
| Use the jointly learned nonlinear 3DMM fitting encoder | Face alignment · AFLW2000 | NME: 5.42 -> 4.70 | Helped |
| Use patchGAN rather than a global image-based discriminator (ablation) | Nonlinear 3DMM texture learning · Rendered face texture | Texture realism and artifacts | Helped |
| Use a global image-based discriminator rather than patchGAN (ablation) | Nonlinear 3DMM texture learning · Rendered face texture | Texture artifacts | Hurt |
Closest papers by meaning
- Dense 3D Face Decoding Over 2500FPS: Joint Texture & Shape Convolutional Mesh Decoders (CVPR 2019 · similarity 0.92)
- MVF-Net: Multi-View 3D Face Morphable Model Regression (CVPR 2019 · similarity 0.92)
- 3D Face Morphable Models "In-The-Wild" (CVPR 2017 · similarity 0.91)
- Uncertainty-Aware Mesh Decoder for High Fidelity 3D Face Reconstruction (CVPR 2020 · similarity 0.91)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.87)
- LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example (CVPR 2024 · similarity 0.86)
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.