Research map / 3D vision / Face and shape
Implicit Neural Head Synthesis via Controllable Local Deformation Fields
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 landmark-centered local deformation fields, attention masks, and local control loss | Implicit Neural Head Synthesis · Subject1 test | LPIPS: 0.0552 -> 0.0465 | Helped |
| use landmark-centered local deformation fields, attention masks, and local control loss | Implicit Neural Head Synthesis · Subject2 test | LPIPS: 0.120 -> 0.0576 | Helped |
| use landmark-centered local deformation fields, attention masks, and local control loss | Implicit Neural Head Synthesis · Subject3 test | LPIPS: 0.052 -> 0.0265 | Helped |
| use landmark-centered local deformation fields, attention masks, and local control loss | Implicit Neural Head Synthesis · Subject4 test | LPIPS: 0.063 -> 0.062 | Helped |
| use the full method rather than the strongest baseline | Implicit Neural Head Synthesis · Subject4 test | ℓ1: 0.065 -> 0.081 | Hurt |
| use the full method rather than the strongest baseline | Implicit Neural Head Synthesis · Subject4 test | SSIM: 0.844 -> 0.830 | Hurt |
| use the full method rather than the strongest baseline | Implicit Neural Head Synthesis · Subject4 test | PSNR: 19.253 -> 19.085 | Hurt |
| add the attention mask to filter expression parameters by local relevance (ablation) | Implicit Neural Head Synthesis · Subject1 test | LPIPS: 0.0487 -> 0.0465 | Helped |
| add the local control loss (ablation) | Implicit Neural Head Synthesis · Subject1 test | LPIPS: 0.0486 -> 0.0465 | Helped |
| replace a global deformation field with multiple local deformation fields (ablation) | Implicit Neural Head Synthesis · Subject1 test | LPIPS: 0.0552 -> 0.0465 | Helped |
| increase local-field landmarks from 5 to 34 (ablation) | Implicit Neural Head Synthesis · test sequences | reconstruction detail | Helped |
Closest papers by meaning
- Efficient 3D Implicit Head Avatar with Mesh-anchored Hash Table Blendshapes (CVPR 2024 · similarity 0.89)
- High-Fidelity 3D Head Avatars Reconstruction through Spatially-Varying Expression Conditioned Neural Radiance Field (AAAI 2024 · similarity 0.89)
- Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians (CVPR 2024 · similarity 0.88)
- DiffusionAvatars: Deferred Diffusion for High-fidelity 3D Head Avatars (CVPR 2024 · similarity 0.88)
- LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example (CVPR 2024 · similarity 0.88)
- 3D Gaussian Head Avatars with Expressive Dynamic Appearances by Compact Tensorial Representations (CVPR 2025 · 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.