SIGGRAPH Asia 2026

UniMate

One Unified Model to Animate Diverse Skeletons

Linzhan Mou, Jiahui Lei, Zhiyang Dou, Chenyue Cai, Chaoyue Song, Adam Finkelstein, Szymon Rusinkiewicz

Given a rigged 3D asset and a text prompt, UniMate generates animations for characters with diverse skeletal topologies within a single unified model.

Interactive Examples.

Drag to orbit Scroll to zoom

View larger

Loading…

Drop .fbx / .glb / .gltf to visualize

Abstract.

Recent advances in automatic rigging now deliver animation-ready 3D assets at scale, yet generating the motion to drive them remains a bottleneck. Existing learned animators are topology-constrained: they rely on category-specific templates or require per-skeleton fine-tuning and reference motions at inference. We present UniMate, a unified foundation model that synthesizes articulated motion for arbitrary skeletons from a rigged 3D asset and a text prompt, with no test-time optimization or per-skeleton retraining. UniMate relies on a topology-aware diffusion transformer, which integrates skeletal topology into attention via three mechanisms: (1) a graph-aware attention bias from pairwise joint relations and geodesic distances; (2) a spectral rotary position embedding generalizing RoPE to arbitrary kinematic trees via the graph Laplacian; and (3) a global topological conditioner attention-pooled from the rest-pose skeleton. We also curate UniML3D, 13,006 motion sequences spanning bipedal, quadrupedal, avian, marine, insectoid, serpentine, and articulated rigid objects with unified canonicalization and text pairing. Trained on this dataset, UniMate outperforms state-of-the-art baselines in quality, generalization, and efficiency, and supports zero-shot cross-topology transfer, in-betweening, expansion, and text-guided editing.

Experiments.

1. Diverse Skeletons, Diverse Prompts.

Diverse skeletons, diverse prompts. One model, no per-skeleton retraining: animals, plants, humanoids, robots, and everyday rigged objects, each animated from the prompt shown above its panel. The three studies below hold one factor fixed at a time.

2. One Prompt, Diverse Skeletons.

One prompt, diverse skeletons. Given a single text prompt, UniMate generates plausible motions across heterogeneous skeletons.

3. One Skeleton, Diverse Prompts.

One skeleton, diverse prompts. Given a single skeleton, UniMate synthesizes diverse motions that follow different input text prompts.

4. One Skeleton, One Prompt, Diverse Motions.

One skeleton, one prompt, diverse motions. Given the same skeleton and text prompt, UniMate generates diverse plausible motion samples.

Applications.

1. Motion Editing.

Motion editing. Given a standing quadruped that shakes its head, we preserve the head motion and regenerate its body and legs to follow the prompt “walks forward.”

2. Motion In-Betweening.

Motion in-betweening. Given the start and end poses and a text prompt, UniMate synthesizes smooth and plausible intermediate motions.

3. Motion Expansion.

Motion expansion. UniMate chains a sequence of text prompts to extend a motion into a longer, temporally coherent sequence.

Dataset.

Training an animator that generalizes across object categories needs motion data with the same breadth, but raw 4D sources are noisy and inconsistent: disconnected or scene-level skeletons, broken roots, non-functional joints, physically implausible motion, and mismatched coordinate frames. We curate UniML3D by unifying three complementary sources — Truebones, Mixamo, and Objaverse-XL.

Sample rigs and motion sequences from UniML3D, spanning bipedal, quadrupedal, avian, marine, insectoid, serpentine, and articulated rigid skeletons.

Samples from UniML3D, spanning bipedal, quadrupedal, avian, marine, insectoid, serpentine, and articulated rigid objects — each with detailed skeleton annotations and a coherent text prompt paired to its motion.

Three passes over every asset. Each line removes (−), adds (+), or re-expresses (→) what the pass was handed.
filter(raw clip) → kept clip
1	prune to a single tree	− every tree but the primary one, by cumulative skinning weight
2	realign the root by forward kinematics	→ a spurious root carried onto the semantic root
3	remove phantom joints	− IK controllers and helper bones with zero skinning weight
4	drop static clips	− sequences with negligible normalized displacement
5	reject implausible clips	− outlier root velocity, or jitter past an anatomical threshold
annotate(kept clip) → captioned clip
6	standardize joint names	→ one anatomical vocabulary, mapped by an LLM
7	resample and render	→ 30 FPS, four synchronized views
8	caption with multimodal LLMs	+ one text prompt per clip
9	review every pair by hand	− captions that misidentify the dominant action
canonicalize(captioned clip) → training sample
10	serialize the tree breadth-first	→ one joint ordering per skeleton
11	scale by topology diameter	→ comparable extent across rigs
12	place in a canonical frame	→ y-axis up, root at the origin
13	align the initial facing direction	→ +z, from the symmetric hip pair
14	express rotations against the rest pose	→ one kinematic basis across topologies
15	normalize by feature statistics	→ global on the root, local on the rest

BibTeX

@inproceedings{mou2026unimate,
  title     = {UniMate: One Unified Model to Animate Diverse Skeletons},
  author    = {Mou, Linzhan and Lei, Jiahui and Dou, Zhiyang and Cai, Chenyue and
               Song, Chaoyue and Finkelstein, Adam and Rusinkiewicz, Szymon},
  booktitle = {SIGGRAPH Asia 2026},
  year      = {2026}
}