Research map / 3D vision / Face and shape
FlexAvatar: Flexible Large Reconstruction Model for Animatable Gaussian Head Avatars with Detailed Deformation
Cheng Peng, Zhuo Su, Liao Wang, Chen Guo, Zhaohu Li, Chengjiang Long, Zheng Lv, Jingxiang Sun, Chenyangguang Zhang, 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 UV position map as driving signal to the UNet (ablation) | FlexAvatar · FaceCap ablation | PSNR: 22.51 -> 23.32 | Improved |
| add UV position map as driving signal to the UNet (ablation) | FlexAvatar · FaceCap ablation | LPIPS: 0.1845 -> 0.1797 | Improved |
| add data distribution adjustment (ablation) | FlexAvatar · FaceCap ablation | PSNR: 22.74 -> 23.32 | Improved |
| add data distribution adjustment (ablation) | FlexAvatar · FaceCap ablation | LPIPS: 0.1868 -> 0.1797 | Improved |
| add mouth-region LPIPS loss (ablation) | FlexAvatar · FaceCap ablation | PSNR: 23.10 -> 23.32 | Improved |
| add mouth-region LPIPS loss (ablation) | FlexAvatar · FaceCap ablation | SSIM: 0.8890 -> 0.8895 | Improved |
| add 10-second test-time refinement to feed-forward model | FlexAvatar · NeRSemble self reenactment | PSNR: 21.1516 -> 22.6313 | Improved |
| add 10-second test-time refinement | FlexAvatar · NeRSemble self reenactment | LPIPS: 0.2193 -> 0.1833 | Improved |
| use FlexAvatar feed-forward instead of best baseline GAGAvatar | FlexAvatar · NeRSemble self reenactment | AKD: 3.8730 -> 3.6518 | Improved |
| use FlexAvatar feed-forward instead of Portrait4D-v2 | FlexAvatar · NeRSemble self reenactment | PSNR: 19.6279 -> 21.1516 | Improved |
| use refined FlexAvatar instead of HeadGAP | FlexAvatar · self reenactment | PSNR: 20.77 -> 22.6313 | Improved |
| add training identities (ablation) | FlexAvatar feed-forward · FaceCap | PSNR | Improved |
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.
- Portrait4D-v1 — Portrait4D: Learning One-Shot 4D Head Avatar Synthesis using Synthetic Data (CVPR 2024) · better on 10 of 10 shared results
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 0.91)
- Generalizable and Animatable Gaussian Head Avatar (NeurIPS 2024 · similarity 0.89)
- RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatars (CVPR 2025 · similarity 0.89)
- HRAvatar: High-Quality and Relightable Gaussian Head Avatar (CVPR 2025 · similarity 0.89)
- FlashAvatar: High-fidelity Head Avatar with Efficient Gaussian Embedding (CVPR 2024 · similarity 0.89)
- GPAvatar: High-fidelity Head Avatars by Learning Efficient Gaussian Projections (CVPR 2025 · similarity 0.88)
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.
