Research map / Video / Video generation
Semantic-Aware Motion Encoding for Topology-Agnostic Character Animation
Zongye Zhang, Yuzhuo Cui, Qingjie Liu, Yunhong Wang
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 spatial transformer (ablation) | SATA autoencoder · AT-HumanML3D recon | JR: 0.1243 -> 0.0568 | Improved |
| add GNN branch (ablation) | SATA autoencoder · AT-HumanML3D recon | JR: 0.0747 -> 0.0568 | Improved |
| add temporal transformer (ablation) | SATA autoencoder · AT-HumanML3D recon | RT: 1.3726 -> 0.9583 | Improved |
| add semantic fusion block (ablation) | SATA autoencoder · AT-HumanML3D recon | JR: 0.1196 -> 0.0568 | Improved |
| add text fusion (ablation) | SATA autoencoder · AT-HumanML3D recon | JR: 0.0663 -> 0.0568 | Improved |
| add sliding window inference (ablation) | SATA VAE · AT-HumanML3D recon | RT: 3.1599 -> 0.9583 | Improved |
| add sliding window inference (ablation) | SATA RVQ · AT-HumanML3D recon | RT: 13.3341 -> 3.5855 | Improved |
| add 16-frame overlap to sliding window (ablation) | SATA VAE · AT-HumanML3D recon | RT: 1.1643 -> 0.9583 | Improved |
| enlarge window from 16/4 to 64/16 (ablation) | SATA VAE · AT-HumanML3D recon | RT: 3.2405 -> 0.9583 | Improved |
| pre-train on AT-HumanML3D before fine-tuning (ablation) | SATA · AT-AniMo4D recon (10% data) | JR: 0.6201 -> 0.2825 | Improved |
| use VAE instead of RVQ bottleneck (ablation) | SATA · AT-HumanML3D retargeting (internal) | retargeting error: 1.1238 -> 0.2122 | Improved |
| use RVQ instead of VAE bottleneck for text-to-motion (ablation) | SATA · AT-HumanML3D text-to-motion | FID: 1.226 -> 0.158 | 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.
- AniMo — AniMo: Species-Aware Model for Text-Driven Animal Motion Generation (CVPR 2025) · 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
- SAMoR: Motion Modelling for Articulated Objects of Any Skeleton and Topology (NeurIPS 2026 · similarity 0.90)
- MotionHiFlow: Text-to-Motion via Hierarchical Flow Matching (CVPR 2026 · similarity 0.88)
- How to Move Your Dragon: Text-to-Motion Synthesis for Large-Vocabulary Objects (ICML 2025 · similarity 0.88)
- OmniMotionGPT: Animal Motion Generation with Limited Data (CVPR 2024 · similarity 0.86)
- AMD: Anatomical Motion Diffusion with Interpretable Motion Decomposition and Fusion (AAAI 2024 · similarity 0.85)
- Single Motion Diffusion (ICLR 2024 · similarity 0.85)
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.
