Research map / 3D vision / Face and shape
Tri^{2}-plane: Thinking Head Avatar via Feature Pyramid
Luchuan Song, Pinxin Liu, Lele Chen, Guojun Yin, Chenliang Xu
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 the 16-patch Phi_512 tri-plane on top of Phi_128 + Phi_256 (full three-level pyramid) (ablation) | Tri^2-plane · Dataset A + B | image quality (LPIPS) | Improved |
| add Phi_256 (4 patches) to a single Phi_128 tri-plane (ablation) | Tri^2-plane · Dataset A + B | F-LMD: 2.35 -> 2.55 | Hurt |
| add Phi_512 (16 patches) to the two-level pyramid (ablation) | Tri^2-plane · Dataset A + B | F-LMD: 2.55 -> 2.48 | Improved |
| add geometry-aware sliding window augmentation (ablation) | Tri^2-plane · Dataset A + B | image quality (rendered maps): Phi_128: 25.50 v.s. 24.70; Phi_256: 8.11 v.s. 5.84 | Improved |
| add geometry-aware sliding window augmentation (ablation) | Tri^2-plane · Dataset A + B | F-LMD: 2.48 v.s. 2.55 | Hurt |
| use Tri^2-plane instead of HAvatar | head avatar · Dataset A + B | PSNR: 28.72 -> 27.75 | Hurt |
| use Tri^2-plane instead of HAvatar | head avatar · Dataset A + B | LPIPS (x10^2): 8.42 -> 5.81 | Improved |
| use Tri^2-plane instead of GaussianHead | head avatar · NeRSemble | PSNR: 29.12 -> 30.04 | Improved |
| use Tri^2-plane instead of HAvatar in user study image quality | head avatar · Self-reenactment user study | MOS1: 4.49 -> 4.43 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PointAvatar — PointAvatar: Deformable Point-Based Head Avatars From Videos (CVPR 2023) · better on 1 of 1 shared result
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
- Generalizable One-shot 3D Neural Head Avatar (NeurIPS 2023 · similarity 0.89)
- Next3D: Generative Neural Texture Rasterization for 3D-Aware Head Avatars (CVPR 2023 · similarity 0.88)
- HyPlaneHead: Rethinking Tri-plane-like Representations in Full-Head Image Synthesis (NeurIPS 2025 · similarity 0.87)
- VOODOO 3D: Volumetric Portrait Disentanglement For One-Shot 3D Head Reenactment (CVPR 2024 · similarity 0.87)
- High-Fidelity 3D Head Avatars Reconstruction through Spatially-Varying Expression Conditioned Neural Radiance Field (AAAI 2024 · similarity 0.87)
- GPAvatar: Generalizable and Precise Head Avatar from Image(s) (ICLR 2024 · similarity 0.87)
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.
