Research map / 3D vision / Human pose estimation
Learning Individual Styles of Conversational Gesture
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 the full speech-to-gesture model instead of the RNN baseline | Speech-to-gesture translation · Test set, all speakers | Avg. L1: 0.70 -> 0.67 | Helped |
| Use the full speech-to-gesture model instead of the RNN baseline | Speech-to-gesture translation · Test set, all speakers | Avg. PCK: 39.69 -> 44.62 | Helped |
| Add the adversarial discriminator (ablation) | Speech-to-gesture model · Test set, all speakers | Avg. L1: 0.67 -> 0.73 | Hurt |
| Add the adversarial discriminator (ablation) | Speech-to-gesture model · Test set, all speakers | Avg. PCK: 44.62 -> 41.95 | Hurt |
| Add speech conditioning to initial-pose conditioning (ablation) | Initial-pose-conditioned gesture model · All speakers | Avg. L1: 0.49 -> 0.47 | Helped |
| Add speech conditioning to initial-pose conditioning (ablation) | Initial-pose-conditioned gesture model · All speakers | Avg. PCK: 61.24 -> 62.39 | Helped |
| Condition on initial pose as well as speech (ablation) | Speech-to-gesture model · All speakers | Avg. L1: 0.67 -> 0.47 | Helped |
| Condition on initial pose as well as speech (ablation) | Speech-to-gesture model · All speakers | Avg. PCK: 44.62 -> 62.39 | Helped |
| Use speech-only conditioning rather than initial-pose-only conditioning (ablation) | Gesture prediction model · All speakers | Avg. L1: 0.49 -> 0.67 | Hurt |
| Use speech-only conditioning rather than initial-pose-only conditioning (ablation) | Gesture prediction model · All speakers | Avg. PCK: 61.24 -> 44.62 | Hurt |
| Use the GAN-based model instead of no-GAN model (ablation) | Gesture prediction for Oliver · Oliver, 4 seconds | Fooled participants (%): 25.0 ± 3.8 -> 35.4 ± 4.0 | Helped |
| Use the GAN-based model instead of no-GAN model (ablation) | Gesture prediction for Meyers · Meyers, 4 seconds | Fooled participants (%): 36.1 ± 4.3 -> 33.2 ± 4.4 | Hurt |
Closest papers by meaning
- ConvoFusion: Multi-Modal Conversational Diffusion for Co-Speech Gesture Synthesis (CVPR 2024 · similarity 0.88)
- Co$^{\mathbf{3}}$Gesture: Towards Coherent Concurrent Co-speech 3D Gesture Generation with Interactive Diffusion (ICLR 2025 · similarity 0.87)
- Weakly-Supervised Emotion Transition Learning for Diverse 3D Co-speech Gesture Generation (CVPR 2024 · similarity 0.86)
- Towards Variable and Coordinated Holistic Co-Speech Motion Generation (CVPR 2024 · similarity 0.86)
- Co-Speech Gesture Video Generation via Motion-Decoupled Diffusion Model (CVPR 2024 · similarity 0.86)
- Emotional Speech-driven 3D Body Animation via Disentangled Latent Diffusion (CVPR 2024 · similarity 0.86)
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.