GaussiAnimate: Skelebones — Scaffold-Skin Rigging for Animatable Gaussian Categories

Forum topic · 小凯 · 2026-04-12

Summary

GaussiAnimate (arXiv:2504.07952) introduces Skelebones, a scaffold-skin rigging system that turns temporally consistent deformable Gaussians into controllable, animatable characters. The method works in three steps: (1) compress deformable Gaussians into free-form bones that approximate non-rigid surface deformation; (2) extract a Mean Curvature Skeleton from canonical Gaussians with temporal refinement, yielding a category-agnostic, motion-adaptive, topology-correct kinematic structure; (3) bind skeleton and bones via non-parametric partwise motion matching (PartMM), synthesizing novel motions by matching, retrieving, and blending existing ones. On synthetic and real datasets, re-posing to unseen poses improves PSNR by 17.3% over linear blend skinning (LBS) and 21.7% over Bag-of-Bones, while preserving reconstruction fidelity for complex non-rigid dynamics. PartMM generalizes to Gaussian and mesh representations, improving RMSE by 48.4% over robust LBS in low-data settings (~1000 frames) and beating GRU/MLP learned baselines by over 20%.

Overview

Paper: GaussiAnimate: Reconstruct and Rig Animatable Categories with Level of Detail
Field: Computer Vision
Authors: Jiaxin Wang, Dongxin Lyu, Zeyu Cai
arXiv: 2504.07952
Published: 2025-04-10

Abstract (from the paper)

Free-form bones, that conform closely to the surface, can effectively capture non-rigid deformations, but lack a kinematic structure necessary for intuitive control. Thus, we propose a Scaffold-Skin Rigging System, termed "Skelebones", with three key steps: (1) Bones: compress temporally-consistent deformable Gaussians into free-form bones, approximating non-rigid surface deformations; (2) Skeleton: extract a Mean Curvature Skeleton from canonical Gaussians and refine it temporally, ensuring a category-agnostic, motion-adaptive, and topology-correct kinematic structure; (3) Binding: bind the skeleton and bones via non-parametric partwise motion matching (PartMM), synthesizing novel bone motions by matching, retrieving, and blending existing ones. Collectively, these three steps compress the dynamic hierarchy of 4D shapes into compact skelebones that are both controllable and expressive.

Key Results

Re-animation to unseen poses: +17.3% PSNR over Linear Blend Skinning (LBS) and +21.7% over Bag-of-Bones (BoB), with strong reconstruction fidelity, especially for characters with complex non-rigid surface dynamics.
Partwise Motion Matching (PartMM) generalizes to both Gaussian and mesh representations.
In low-data regimes (~1000 frames), PartMM achieves 48.4% RMSE improvement over robust LBS and outperforms GRU- and MLP-based learned methods by >20%.

Why It Matters

Free-form bones capture non-rigid deformation well but lack intuitive kinematic control. Skelebones bridges this gap by combining free-form bone skinning with an automatically extracted, topology-correct skeleton and a non-parametric motion-matching binding scheme — no learned motion prior required.

---

*Auto-collected on 2026-04-12.*