Research map / 3D vision / Face and shape
Motion-Aware Animatable Gaussian Avatars Deblurring
Muyao Niu, Yifan Zhan, Qingtian Zhu, Zhuoxiao Li, Wei Wang, Zhihang Zhong, Xiao Sun, Yinqiang Zheng
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 B-spline pose interpolation over per-timestep pose optimization (ablation) | Ours · Synthetic ZJU-MoCap blur (K_blur=5, 4 views) | PSNR: 24.009 -> 25.546 | Improved |
| add pose deformation model (ablation) | Ours · Synthetic ZJU-MoCap blur (K_blur=5, 4 views) | PSNR: 25.301 -> 25.546 | Improved |
| add LBS weight optimization (ablation) | Ours · Synthetic ZJU-MoCap blur (K_blur=5, 4 views) | PSNR: 25.394 -> 25.546 | Improved |
| add SMPL shape optimization (ablation) | Ours · Synthetic ZJU-MoCap blur (K_blur=5, 4 views) | PSNR: 25.529 -> 25.546 | Improved |
| add pose deformation model (ablation) | Ours · Real hybrid-exposure dataset | PSNR: 26.426 -> 27.010 | Improved |
| add inter-frame motion regularization L_reg (K_blur=5) (ablation) | Ours · Synthetic non-middle timesteps | PSNR: 24.421 -> 25.417 | Improved |
| add inter-frame motion regularization L_reg (K_blur=5) (ablation) | Ours · Synthetic middle timestep | PSNR: 25.567 -> 25.546 | No clear effect |
| add SMPL parameter optimization (pose, shape, LBS) (ablation) | Ours · Synthetic ZJU-MoCap blur (K_blur=5, 4 views) | PSNR: 21.63 -> 25.55 | Improved |
| use B-spline instead of linear trajectory (ablation) | Ours · Synthetic ZJU-MoCap blur (K_blur=5, 4 views) | PSNR: 25.516 -> 25.546 | Improved |
| use B-spline instead of linear trajectory (ablation) | Ours · Real hybrid-exposure dataset | PSNR: 26.899 -> 27.010 | Improved |
| increase control knots P from 2 to 4 (ablation) | Ours · Real hybrid-exposure dataset | PSNR: 26.899 -> 27.010 | Improved |
| increase virtual sharp images T from 3 to 5 (ablation) | Ours · Real hybrid-exposure dataset | PSNR: 26.837 -> 27.010 | 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.
- GauHuman — GauHuman: Articulated Gaussian Splatting from Monocular Human Videos (CVPR 2024) · better on 8 of 8 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
- GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians (CVPR 2024 · similarity 0.89)
- 3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting (CVPR 2024 · similarity 0.89)
- Real-time High-fidelity Gaussian Human Avatars with Position-based Interpolation of Spatially Distributed MLPs (CVPR 2025 · similarity 0.88)
- AniGS: Animatable Gaussian Avatar from a Single Image with Inconsistent Gaussian Reconstruction (CVPR 2025 · similarity 0.88)
- Human Gaussian Splatting: Real-time Rendering of Animatable Avatars (CVPR 2024 · similarity 0.88)
- 3D²-Actor: Learning Pose-Conditioned 3D-Aware Denoiser for Realistic Gaussian Avatar Modeling (AAAI 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.
