Research map / 3D vision / Face and shape
GeoDiff4D: Geometry-Aware Diffusion for 4D Head Avatar Reconstruction
Chao Xu, Xiaochen Zhao, Xiang Deng, Jingxiang Sun, Donglin Di, Zhuo Su, Yebin Liu
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 cross-view pairing to expression encoder training (ablation) | GeoDiff4D VGM · NeRSemblev2 self-reenactment | PSNR: 19.895 -> 21.586 | Improved |
| add joint image-normal representation (ablation) | GeoDiff4D VGM · NeRSemblev2 self-reenactment | PSNR: 20.809 -> 21.586 | Improved |
| add Domain-Spatial attention (ablation) | GeoDiff4D VGM · NeRSemblev2 self-reenactment | PSNR: 20.984 -> 21.586 | Improved |
| add synthetic SynthHuman data to training (ablation) | GeoDiff4D VGM · NeRSemblev2 self-reenactment | PSNR: 20.892 -> 21.586 | Improved |
| add hierarchical refinement in avatar reconstruction (ablation) | GeoDiff4D reconstruction · NeRSemblev2 self-reenactment | PSNR: 19.816 -> 19.953 | Improved |
| add surface normal regularization (ablation) | GeoDiff4D reconstruction · NeRSemblev2 self-reenactment | AED: 2.713 -> 2.563 | Improved |
| use video-generated normals instead of DAViD monocular normals (ablation) | GeoDiff4D reconstruction · NeRSemblev2 self-reenactment | PSNR: 19.947 -> 19.953 | No clear effect |
| GeoDiff4D versus CAP4D baseline | GeoDiff4D · NeRSemblev2 self-reenactment | PSNR: 19.295 -> 19.951 | 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.
- CAP4D — CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models (CVPR 2025) · better on 6 of 6 shared results
- Portrait4D-v2 — Portrait4D: Learning One-Shot 4D Head Avatar Synthesis using Synthetic Data (CVPR 2024) · 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
- GAF: Gaussian Avatar Reconstruction from Monocular Videos via Multi-view Diffusion (CVPR 2025 · similarity 0.91)
- CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models (CVPR 2025 · similarity 0.91)
- High-Fidelity 3D Head Avatars Reconstruction through Spatially-Varying Expression Conditioned Neural Radiance Field (AAAI 2024 · similarity 0.91)
- DiffusionAvatars: Deferred Diffusion for High-fidelity 3D Head Avatars (CVPR 2024 · similarity 0.91)
- Human-3Diffusion: Realistic Avatar Creation via Explicit 3D Consistent Diffusion Models (NeurIPS 2024 · similarity 0.90)
- 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.
