Research map / 3D vision / Face and shape
STAvatar: Soft Binding and Temporal Density Control for Monocular 3D Head Avatars Reconstruction
Jiankuo Zhao, Xiangyu Zhu, Zidu Wang, Zhen Lei
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 UV-Adaptive Soft Binding (ablation) | STAvatar · INSTA | PSNR: 29.66 -> 30.63 | Improved |
| add adaptive density control (ablation) | STAvatar · INSTA | LPIPS: 0.0505 -> 0.0304 | Improved |
| add adaptive density control (ablation) | STAvatar · INSTA | PSNR: 30.31 -> 30.63 | Improved |
| replace positional gradient with FPE-AP (ablation) | STAvatar · INSTA | LPIPS: 0.0321 -> 0.0304 | Improved |
| add FLAME-conditioned temporal clustering (ablation) | STAvatar · INSTA | LPIPS: 0.0310 -> 0.0304 | Improved |
| add local decoder branch (ablation) | STAvatar · INSTA | PSNR: 30.36 -> 30.63 | Improved |
| add VGG perceptual loss (ablation) | STAvatar · INSTA | LPIPS: 0.0426 -> 0.0304 | Improved |
| add VGG perceptual loss (ablation) | STAvatar · INSTA | SSIM: 0.9592 -> 0.9587 | Hurt |
| apply FTC for densification (ablation) | STAvatar · INSTA mouth-interior Gaussian count | Gaussian count: 2337 -> 2751 | Improved |
| use STAvatar instead of prior Gaussian avatar methods | Gaussian head avatars · PointAvatar / HDTF | PSNR: 28.25 vs 28.36 (Fate); 27.99 vs 28.36 | Mixed |
Closest papers by meaning
- HRAvatar: High-Quality and Relightable Gaussian Head Avatar (CVPR 2025 · similarity 0.91)
- RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatars (CVPR 2025 · similarity 0.91)
- GPAvatar: High-fidelity Head Avatars by Learning Efficient Gaussian Projections (CVPR 2025 · similarity 0.90)
- FATE: Full-head Gaussian Avatar with Textural Editing from Monocular Video (CVPR 2025 · similarity 0.89)
- Synthetic Prior for Few-Shot Drivable Head Avatar Inversion (CVPR 2025 · similarity 0.89)
- Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians (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.
