Research map / Video / Video generation
ReactDance: Hierarchical Representation for High-Fidelity and Coherent Long-Form Reactive Dance Generation
Jingzhong Lin, Xinru Li, Yuanyuan Qi, Bohao Zhang, Wenxiang Liu, Kecheng Tang, Wenxuan Huang, Xiangfeng Xu, Bangyan Li, Changbo Wang, Gaoqi He
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 |
|---|---|---|---|
| use HFSQ instead of RVQ-VAE tokenizer (with PM) (ablation) | ReactDance · DD100 generation | FID_g: 26.98 -> 7.63 | Improved |
| use HFSQ instead of plain VAE (ablation) | ReactDance · DD100 reconstruction | MPJPE: 40.87 -> 28.66 | Improved |
| add progressive masking to HFSQ (ablation) | ReactDance · DD100 generation | FID_g: 10.46 -> 7.63 | Improved |
| add progressive masking to HFSQ (ablation) | ReactDance · DD100 reconstruction | MPJPE: 24.15 -> 28.66 | Hurt |
| reduce training stride from 16 to 4 (dense sliding window) (ablation) | ReactDance · DD100 duet | FID_cd: 22.69 -> 14.17 | Improved |
| reduce training stride from 64 to 4 (ablation) | ReactDance · DD100 duet | FID_cd: 39.50 -> 14.17 | Improved |
| use periodic encoding and causal masking instead of latent stitching (ablation) | ReactDance · DD100 duet | FID_cd: 18.59 -> 14.17 | Improved |
| use periodic encoding and causal masking instead of latent stitching (ablation) | ReactDance · DD100 | AITS: 2.03 -> 1.75 | Improved |
| use R=2 residual stages instead of R=1 (ablation) | HFSQ · DD100 generation | FID_k: 13.10 -> 5.57 | Improved |
| increase residual stages from R=2 to R=4 (ablation) | HFSQ · DD100 generation | FID_k: 5.57 -> 6.10 | Hurt |
| add leader motion condition (ablation) | ReactDance · DD100 duet | FID_cd: 48.82 -> 14.17 | Improved |
| add music condition (ablation) | ReactDance · DD100 duet | FID_cd: 35.35 -> 14.17 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- GestureLSM — GestureLSM: Latent Shortcut based Co-Speech Gesture Generation with Spatial-Temporal Modeling (ICCV 2025) · better on 4 of 4 shared results
- InterGen — InterGen: Diffusion-Based Multi-human Motion Generation Under Complex Interactions (IJCV 2024) · better on 4 of 4 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 0.91)
- MEGADance: Mixture-of-Experts Architecture for Genre-Aware 3D Dance Generation (NeurIPS 2025 · similarity 0.90)
- DanceAnyWay: Synthesizing Beat-Guided 3D Dances with Randomized Temporal Contrastive Learning (AAAI 2024 · similarity 0.90)
- Bidirectional Autoregessive Diffusion Model for Dance Generation (CVPR 2024 · similarity 0.89)
- Breaking the Group Size Barrier: Parameter-Efficient Group Dance Generation with Chain-of-Dancers (NeurIPS 2026 · similarity 0.89)
- Lodge: A Coarse to Fine Diffusion Network for Long Dance Generation Guided by the Characteristic Dance Primitives (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.
