Research map / 3D vision / Face and shape
Dense 3D Face Decoding Over 2500FPS: Joint Texture & Shape Convolutional Mesh Decoders
Check 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 |
|---|---|---|---|
| use CMD instead of 3DDFA | 3D face alignment · AFLW2000-3D | NME (%): 5.42 -> 3.98 | Helped |
| use CMD instead of N-3DMM | 3D face alignment · AFLW2000-3D | NME (%): 4.12 -> 3.98 | Helped |
| use CMD instead of PRNet | 3D face alignment · AFLW2000-3D | NME (%): 3.62 -> 3.98 | Hurt |
| add joint texture decoding to shape-only mesh decoding (ablation) | Mesh decoder, latent dimension 64 · 3DMD test set | Shape NME: 0.0304 -> 0.0286 | Helped |
| add joint texture decoding to shape-only mesh decoding (ablation) | Mesh decoder, latent dimension 128 · 3DMD test set | Shape NME: 0.0261 -> 0.0220 | Helped |
| use joint nonlinear shape-texture decoding instead of PCA (ablation) | 3D face model, dimension 64 · 3DMD test set | Shape NME: 0.0313 -> 0.0286 | Helped |
| use joint nonlinear shape-texture decoding instead of PCA (ablation) | 3D face model, dimension 128 · 3DMD test set | Shape NME: 0.0280 -> 0.0220 | Helped |
| use joint nonlinear shape-texture decoding instead of PCA (ablation) | 3D face model, dimension 64 · 3DMD test set | Texture l1 channel-wise error: 0.0196 -> 0.0325 | Hurt |
| increase joint decoder embedding dimension from 64 to 256 (ablation) | CMD · 3DMD test set | Shape NME: 0.0286 -> 0.0133 | Helped |
| increase joint decoder embedding dimension from 64 to 256 (ablation) | CMD · 3DMD test set | Texture l1 channel-wise error: 0.0325 -> 0.0228 | Helped |
| use CMD instead of N-3DMM for mesh decoding | 3D face reconstruction · Runtime comparison | Decoder time: 5.5 ms -> 0.367 ms | Helped |
Closest papers by meaning
- Nonlinear 3D Face Morphable Model (CVPR 2018 · similarity 0.92)
- Uncertainty-Aware Mesh Decoder for High Fidelity 3D Face Reconstruction (CVPR 2020 · similarity 0.91)
- Regressing Robust and Discriminative 3D Morphable Models With a Very Deep Neural Network (CVPR 2017 · similarity 0.89)
- 3D Face Morphable Models "In-The-Wild" (CVPR 2017 · similarity 0.89)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.86)
- Locally Adaptive Neural 3D Morphable Models (CVPR 2024 · similarity 0.85)
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.