Research map / Video / Video segmentation and tracking
How Does it Sound?
Kun Su, Xiulong Liu, Eli Shlizerman
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 ST-GCN to transformer for beat prediction (ablation) | Video2Rhythm · AIST | CMLc: 16.07 -> 54.89 | Improved |
| add transformer on top of ST-GCN (ablation) | Video2Rhythm · AIST | F-score: 64.78 -> 71.93 | Improved |
| add self-similarity matrix (SSM) (ablation) | ST-GCN+TF · AIST | F-score: 71.93 -> 73.07 | Improved |
| add relative attention (ablation) | ST-GCN+TF · AIST | F-score: 71.93 -> 74.67 | Improved |
| combine SSM and relative attention (ablation) | ST-GCN+TF · AIST | CMLc: 61.89 -> 71.43 (+26.46%) | Improved |
| full Video2Rhythm instead of TCN baseline (ablation) | Video2Rhythm · AIST | F-score: 63.04 -> 75.79 | Improved |
| generate hits as a word-token sequence (ablation) | Rhythm2Drum transformer · Groove MIDI | NDB: 44 -> 39 | Improved |
| add hits sequence tokens (ablation) | Rhythm2Drum transformer · Groove MIDI | MSE Velocity: 0.0507 -> 0.0493 | Mixed |
| add U-net for velocities and offsets (ablation) | Rhythm2Drum · Groove MIDI | MSE Offsets: 0.0369 -> 0.0169 (↓58%) | Improved |
| full Rhythm2Drum vs GrooveVAE (ablation) | Rhythm2Drum · Groove MIDI | MSE Velocity: 0.0437 -> 0.0267 (↓40%) | Improved |
| add recurrence memory (ablation) | Drum2Piano · Lakh MIDI | validation NLL: 0.77 -> 0.53 | Improved |
| add recurrence memory (ablation) | Drum2Guitar · Lakh MIDI | validation NLL: 0.58 -> 0.40 | Improved |
Closest papers by meaning
- Long-Term Rhythmic Video Soundtracker (ICML 2023 · similarity 0.89)
- OmniDance: Multimodal Driven Dance Video Generation with Large-scale Internet Data (ECCV 2026 · similarity 0.88)
- Music-Aligned Holistic 3D Dance Generation via Hierarchical Motion Modeling (ICCV 2025 · similarity 0.87)
- TMD-Bench: A Multi-Level Evaluation Paradigm for Music–Dance Co-Generation (ICML 2026 · similarity 0.87)
- DanceAnyWay: Synthesizing Beat-Guided 3D Dances with Randomized Temporal Contrastive Learning (AAAI 2024 · similarity 0.87)
- BRACE: The Breakdancing Competition Dataset for Dance Motion Synthesis (ECCV 2022 · similarity 0.87)
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.
