Research map / 3D vision / Face and shape
Fresco: Frequency-Spatial Consistent Optimization for Fine-Grained Head Avatar Modeling
Shikun Zhang, Yong Li, Yiqun Wang, Qiuhong Ke, Cunjian Chen
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 high-frequency enhancement loss (DoG edge-aware + gradient difference) (ablation) | Fresco · NeRSemble subject 253, novel-view | PSNR: 32.60 -> 34.02 | Improved |
| add UV-space cross-view consistency loss (ablation) | Fresco · NeRSemble subject 253, novel-view | PSNR: 32.71 -> 34.02 | Improved |
| add UV-space cross-view consistency loss (ablation) | Fresco · NeRSemble subject 253, self-reenactment | PSNR: 31.47 -> 32.46 | Improved |
| add low-frequency regularization with decaying weight (ablation) | Fresco · NeRSemble subject 253, novel-view | PSNR: 33.48 -> 34.02 | Improved |
| add seam regularization (ablation) | Fresco · NeRSemble subject 253, novel-view | PSNR: 33.86 -> 34.02 | No clear effect |
| add seam regularization (ablation) | Fresco · NeRSemble subject 253, qualitative seams | visual continuity | Improved |
| use Fresco instead of GaussianAvatars | novel-view synthesis · NeRSemble novel-view synthesis (8 subjects) | PSNR: 31.50 -> 31.80 | Improved |
| use Fresco instead of GaussianAvatars | self-reenactment · NeRSemble self-reenactment (8 subjects) | SSIM: 0.930 -> 0.929 | No clear effect |
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.
- GaussianAvatars — GaussianAvatars: Photorealistic Head Avatars with Rigged 3D Gaussians (CVPR 2024) · better on 5 of 6 shared results
- Gaussian Head Avatar — Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians (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
- Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians (CVPR 2024 · similarity 0.87)
- High-Fidelity 3D Head Avatars Reconstruction through Spatially-Varying Expression Conditioned Neural Radiance Field (AAAI 2024 · similarity 0.87)
- MeGA: Hybrid Mesh-Gaussian Head Avatar for High-Fidelity Rendering and Head Editing (CVPR 2025 · similarity 0.87)
- RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatars (CVPR 2025 · similarity 0.87)
- FATE: Full-head Gaussian Avatar with Textural Editing from Monocular Video (CVPR 2025 · similarity 0.86)
- FA-LAM: Focus-Aware Large Avatar Model for One-Shot 4D Animatable Gaussian Head (NeurIPS 2026 · similarity 0.86)
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.
