Research map / Vision-language and multimodal / Cross-modal learning
Learning Hierarchical Cross-Modal Association for Co-Speech Gesture Generation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 hierarchical pose generation instead of holistic pose generation (ablation) | HA2G · TED-Expressive | FGD: 11.989 -> 5.306 | Helped |
| Add hand hierarchy (ablation) | HA2G · TED-Expressive | FGD: 10.832 -> 5.306 | Helped |
| Add body hierarchy (ablation) | HA2G · TED-Expressive | FGD: 5.882 -> 5.306 | Helped |
| Use distinct audio features for each pose hierarchy level (ablation) | HA2G · TED-Expressive | FGD: 6.801 -> 5.306 | Helped |
| Add physical pose regularization (ablation) | HA2G · TED-Expressive | FGD: 5.907 -> 5.306 | Helped |
| Use multiple audio-feature levels instead of only low-level features (ablation) | HA2G · TED-Expressive | FGD: 6.588 -> 5.306 | Helped |
| Use multiple audio-feature levels instead of only mid-level features (ablation) | HA2G · TED-Expressive | FGD: 7.212 -> 5.306 | Helped |
| Use multiple audio-feature levels instead of only high-level features (ablation) | HA2G · TED-Expressive | FGD: 7.421 -> 5.306 | Helped |
| Add high-level audio negatives to contrastive learning (ablation) | HA2G · TED-Expressive | FGD: 7.982 -> 5.306 | Helped |
| Add low- and mid-level audio negatives to contrastive learning (ablation) | HA2G · TED-Expressive | FGD: 6.998 -> 5.306 | Helped |
| Add transcript-guided audio-text contrastive learning (ablation) | HA2G · TED-Expressive | FGD: 9.228 -> 5.306 | Helped |
| Replace contrastive audio learning with a pretrained ASR encoder (ablation) | HA2G · TED-Expressive | FGD: 5.319 vs 5.306 | No clear effect |
Closest papers by meaning
- Co$^{\mathbf{3}}$Gesture: Towards Coherent Concurrent Co-speech 3D Gesture Generation with Interactive Diffusion (ICLR 2025 · similarity 0.91)
- Co-Speech Gesture Video Generation with Implicit Motion-Audio Entanglement (CVPR 2025 · similarity 0.90)
- ConvoFusion: Multi-Modal Conversational Diffusion for Co-Speech Gesture Synthesis (CVPR 2024 · similarity 0.89)
- Co-Speech Gesture Video Generation via Motion-Decoupled Diffusion Model (CVPR 2024 · similarity 0.88)
- Towards Variable and Coordinated Holistic Co-Speech Motion Generation (CVPR 2024 · similarity 0.88)
- HOP: Heterogeneous Topology-based Multimodal Entanglement for Co-Speech Gesture Generation (CVPR 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.