Research map / 3D vision / Face and shape
Learning Personalized High Quality Volumetric Head Avatars From Monocular RGB Videos
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 |
|---|---|---|---|
| Use the 3DMM-displacement-conditioned U-Net feature predictor instead of the code-conditioned CNN decoder (ablation) | Ours-D · Out-of-training expressions | expression robustness | Helped |
| Use UV-space CNN feature prediction instead of static features (ablation) | 3DMM-anchored NeRF · Subject0 | LPIPS: 0.1559 -> 0.1417 | Helped |
| Add 3DMM expression and pose codes to static vertex features (ablation) | 3DMM-anchored NeRF · Subject0 | LPIPS: 0.1559 -> 0.1599 | Hurt |
| Replace shallow code-conditioned feature decoding with a code-conditioned MLP (ablation) | 3DMM-anchored NeRF · Subject0 | LPIPS: 0.1599 -> 0.1568 | No clear effect |
| Predict UV features from 3DMM codes with a CNN decoder rather than a code-conditioned MLP (ablation) | 3DMM-anchored NeRF · Subject0 | LPIPS: 0.1568 -> 0.1417 | Helped |
| Use 3DMM vertex displacements as feature-predictor input rather than 3DMM codes (ablation) | Ours-D · Out-of-training expressions | expression robustness | Helped |
| Reduce training data to the first 50% of frames (ablation) | Ours-C · Subject0 | LPIPS: 0.1417 -> 0.2038 | Hurt |
| Reduce training data to the first 50% of frames (ablation) | Ours-D · Subject0 | LPIPS: 0.1439 -> 0.1711 | Hurt |
| Use Ours-D rather than Ours-C with reduced training data (ablation) | UV-space feature predictors · Subject0, first 50% training data | LPIPS: 0.2038 -> 0.1711 | Helped |
| Use the proposed Ours-D avatar rather than NHA | Monocular head avatar · Subject0 | LPIPS: 0.165 -> 0.144 | Helped |
Closest papers by meaning
- High-Fidelity 3D Head Avatars Reconstruction through Spatially-Varying Expression Conditioned Neural Radiance Field (AAAI 2024 · similarity 0.90)
- Efficient 3D Implicit Head Avatar with Mesh-anchored Hash Table Blendshapes (CVPR 2024 · similarity 0.90)
- HRAvatar: High-Quality and Relightable Gaussian Head Avatar (CVPR 2025 · similarity 0.90)
- RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatars (CVPR 2025 · similarity 0.89)
- Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians (CVPR 2024 · similarity 0.89)
- DiffusionAvatars: Deferred Diffusion for High-fidelity 3D Head Avatars (CVPR 2024 · similarity 0.89)
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.