Research map / Video / Video generation
MDD: A Dataset for Text-and-Music Conditioned Duet Dance Generation
Prerit Gupta, Jason Alexander Fotso-Puepi, Zhengyuan Li, Jay Mehta, Aniket Bera
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 InterGen instead of MDM (both modalities) | Text-to-Duet baselines · MDD Text-to-Duet | FID: 1.739 -> 0.426 | Improved |
| add music to text-only InterGen (ablation) | InterGen · MDD Text-to-Duet | FID: 0.405 -> 0.426 | Hurt |
| add text to music-only InterGen (ablation) | InterGen · MDD Text-to-Duet | FID: 2.014 -> 0.426 | Improved |
| use Jukebox music features instead of MFCC (ablation) | InterGen (both) · MDD Text-to-Duet | R-Precision Top-1: 0.105 -> 0.138 | Improved |
| use Jukebox music features instead of MFCC (ablation) | InterGen (both) · MDD Text-to-Duet | BED: 0.385 -> 0.454 | Improved |
| add music to text-only DuoLando (ablation) | DuoLando · MDD Text-to-Dance Accompaniment | FID: 1.538 -> 0.698 | Improved |
| add text to music-only DuoLando (ablation) | DuoLando · MDD Text-to-Dance Accompaniment | FID: 0.721 -> 0.698 | Improved |
| add text to music-only DuoLando (ablation) | DuoLando · MDD Text-to-Dance Accompaniment | BED: 0.305 -> 0.395 | Improved |
| use GPT-4o refined text instead of raw text (ablation) | InterGen · MDD Text-to-Duet | FID: 0.511 -> 0.426 | Improved |
| use GPT-4o refined text instead of move names (ablation) | InterGen · MDD Text-to-Duet | FID: 0.721 -> 0.426 | Improved |
| use GPT-4o refined text instead of raw text (ablation) | InterGen · MDD Text-to-Duet | MM Dist: 1.722 -> 1.532 | Improved |
| add text to music-only MDM (ablation) | MDM · MDD Text-to-Duet | BAS: 0.192 -> 0.231 | Mixed |
Closest papers by meaning
- TM2D: Bimodality Driven 3D Dance Generation via Music-Text Integration (ICCV 2023 · similarity 0.92)
- OpenDance: Multimodal Controllable 3D Dance Generation with Large-scale Internet Data (CVPR 2026 · similarity 0.92)
- OmniDance: Multimodal Driven Dance Video Generation with Large-scale Internet Data (ECCV 2026 · similarity 0.91)
- DanceAnyWay: Synthesizing Beat-Guided 3D Dances with Randomized Temporal Contrastive Learning (AAAI 2024 · similarity 0.89)
- Duolando: Follower GPT with Off-Policy Reinforcement Learning for Dance Accompaniment (ICLR 2024 · similarity 0.88)
- MEGADance: Mixture-of-Experts Architecture for Genre-Aware 3D Dance Generation (NeurIPS 2025 · 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.
