Research map / 3D vision / 3D generation
InfiniteDance: Scalable 3D Dance Generation Towards in-the-wild Generalization
Ronghui Li, zhongyuan hu, Li Siyao, youliang zhang, Haozhe Xie, Mingyuan Zhang, Jie Guo, Xiu Li, Ziwei 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 |
|---|---|---|---|
| feed continuous token embeds instead of discrete token indices (ablation) | ChoreoLLaMA · InfiniteDance test | FID_k: 79.84 -> 62.87 | Improved |
| feed continuous token embeds instead of discrete token indices (ablation) | ChoreoLLaMA · InfiniteDance test | Div_k: 13.09 -> 5.49 | Hurt |
| add RAG reference dance retrieval (ablation) | ChoreoLLaMA · InfiniteDance test | FID_k: 62.87 -> 38.74 | Improved |
| add Cadence-MoE on top of RAG (ablation) | ChoreoLLaMA · InfiniteDance test | FID_k: 38.74 -> 30.54 | Improved |
| add Cadence-MoE on top of RAG (ablation) | ChoreoLLaMA · InfiniteDance test | BAS: 0.2325 -> 0.2342 | Improved |
| replace Lodge with ChoreoLLaMA | dance generation · FineDance (cross-dataset) | FID_k: 106.85 -> 59.38 | Improved |
| replace Lodge with ChoreoLLaMA | dance generation · Unseen Music (OOD) | BAS: 0.2332 -> 0.2315 | Hurt |
| replace Lodge with ChoreoLLaMA | dance generation · Unseen Music (OOD) | FID_k: 119.66 -> 56.22 | Improved |
| add FRDM foot restoration after physics imitation (vs PHC only) (ablation) | motion reconstruction · in-the-wild dance videos | foot skating ratio: 8.87% -> 5.09% | Improved |
| add FRDM instead of simple smoothing after PHC (ablation) | motion reconstruction · in-the-wild dance videos | foot skating ratio: 14.29% -> 5.09% | Improved |
| add physics imitation (PHC) to GVHMR (ablation) | motion reconstruction · in-the-wild dance videos | jitter: 31.89 -> 78.60 | Hurt |
| add physics imitation (PHC) to GVHMR (ablation) | motion reconstruction · in-the-wild dance videos | foot skating ratio: 28.63% -> 8.87% | 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.
- Bailando — Bailando: 3D Dance Generation by Actor-Critic GPT With Choreographic Memory (CVPR 2022) · better on 2 of 2 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
- OpenDance: Multimodal Controllable 3D Dance Generation with Large-scale Internet Data (CVPR 2026 · similarity 0.93)
- MEGADance: Mixture-of-Experts Architecture for Genre-Aware 3D Dance Generation (NeurIPS 2025 · similarity 0.92)
- DanceAnyWay: Synthesizing Beat-Guided 3D Dances with Randomized Temporal Contrastive Learning (AAAI 2024 · similarity 0.90)
- DanceCamera3D: 3D Camera Movement Synthesis with Music and Dance (CVPR 2024 · similarity 0.90)
- POPDG: Popular 3D Dance Generation with PopDanceSet (CVPR 2024 · similarity 0.89)
- Bidirectional Autoregessive Diffusion Model for Dance Generation (CVPR 2024 · 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.
