Research map / 3D vision / 3D generation
FlexAvatar: Learning Complete 3D Head Avatars with Partial Supervision
Tobias Kirschstein, Simon Giebenhain, Matthias Nießner
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 bias sinks when mixing monocular and multi-view data (ablation) | FlexAvatar · Ava256 (single-image, ablation) | PSNR: 14.5 -> 17.2 | Improved |
| add bias sinks when mixing monocular and multi-view data (ablation) | FlexAvatar · Ava256 (single-image, ablation) | LPIPS: 0.351 -> 0.285 | Improved |
| add multi-view 3D data to monocular-only training (ablation) | FlexAvatar · Ava256 (single-image, ablation) | PSNR: 13.7 -> 17.2 | Improved |
| add monocular 2D data to multi-view-only training (ablation) | FlexAvatar · Ava256 (single-image, ablation) | CSIM: 0.119 -> 0.621 | Improved |
| add StyleGAN-PixelShuffle upsampler (ablation) | FlexAvatar · Ava256 (single-image, ablation) | PSNR: 17.1 -> 17.2 | Improved |
| add StyleGAN-PixelShuffle upsampler (ablation) | FlexAvatar · Ava256 (single-image, ablation) | AKD: 7.03 -> 6.34 | Improved |
| add per-person fitting of avatar code (ablation) | FlexAvatar · Ava256 (single-image, ablation) | CSIM: 0.621 -> 0.682 | Improved |
| add per-person fitting of avatar code (ablation) | FlexAvatar · Ava256 (single-image, ablation) | PSNR: 17.2 -> 16.9 | Hurt |
| use FlexAvatar instead of LAM | LAM · VFHQ-Test (self-reenactment) | AKD: 2.059 -> 2.965 | Hurt |
| use FlexAvatar instead of LAM (single-image) | LAM · Ava256 (single-image) | PSNR: 13.1 -> 16.9 | Improved |
| use FlexAvatar instead of Avat3r (few-shot) | Avat3r · Ava256 (few-shot, 4 images) | LPIPS: 0.310 -> 0.218 | Improved |
| use FlexAvatar instead of RGBAvatar | RGBAvatar · NeRSemble monocular benchmark | LPIPS: 0.181 -> 0.156 | Improved |
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.
- RGBAvatar — RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatars (CVPR 2025) · better on 3 of 3 shared results
- Avat3r — Avat3r: Large Animatable Gaussian Reconstruction Model for High-fidelity 3D Head Avatars (ICCV 2025) · better on 3 of 3 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
- Arc2Avatar: Generating Expressive 3D Avatars from a Single Image via ID Guidance (CVPR 2025 · similarity 0.89)
- HRAvatar: High-Quality and Relightable Gaussian Head Avatar (CVPR 2025 · similarity 0.89)
- FATE: Full-head Gaussian Avatar with Textural Editing from Monocular Video (CVPR 2025 · similarity 0.88)
- Synthetic Prior for Few-Shot Drivable Head Avatar Inversion (CVPR 2025 · similarity 0.88)
- GeneAvatar: Generic Expression-Aware Volumetric Head Avatar Editing from a Single Image (CVPR 2024 · similarity 0.88)
- GPAvatar: Generalizable and Precise Head Avatar from Image(s) (ICLR 2024 · similarity 0.88)
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.
