Research map / Video / Video generation
Democratizing High-Fidelity Co-Speech Gesture Video Generation
Xu Yang, Shaoli Huang, Shenbo Xie, Xuelin Chen, Yifei Liu, Changxing Ding
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 reference skeleton input to audio-to-skeleton model (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (unpaired) | FID: 127.40 -> 62.93 | Improved |
| add reference skeleton input to audio-to-skeleton model (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (paired) | PSNR: 13.75 -> 16.62 | Improved |
| use feature concatenation for audio instead of cross-attention (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (unpaired) | Sync-C: 0.77 -> 6.28 | Improved |
| use feature concatenation for audio instead of cross-attention (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (unpaired) | Sync-D: 13.64 -> 8.68 | Improved |
| use feature concatenation for audio instead of cross-attention (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (unpaired) | FID: 78.10 -> 62.93 | Improved |
| add local motion representation (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (unpaired) | Sync-C: 0.49 -> 6.28 | Improved |
| add local motion representation (ablation) | Ours (StableAnimator, fine-tuned) · CSG-405 (unpaired) | FID: 86.69 -> 62.93 | Improved |
| fine-tune StableAnimator on CSG-405 | Ours (StableAnimator) · CSG-405 (unpaired) | FID: 66.47 -> 62.93 | Improved |
| fine-tune StableAnimator on CSG-405 | Ours (StableAnimator) · PATS (unpaired) | FID: 73.56 -> 70.38 | Improved |
| fine-tune StableAnimator on CSG-405 | Ours (StableAnimator) · TED-talks (unpaired) | FID: 68.84 -> 70.05 | Hurt |
| generate skeleton from audio instead of ground-truth skeleton | StableAnimator · CSG-405 (paired) | SSIM: 0.83 -> 0.65 | Hurt |
| generate body-shape-aware skeletons instead of skeletons from another video | StableAnimator · CSG-405 (unpaired) | FID: 94.94 -> 66.47 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MimicMotion — MimicMotion: High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance (ICML 2025) · better on 3 of 3 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
- Co-Speech Gesture Video Generation via Motion-Decoupled Diffusion Model (CVPR 2024 · similarity 0.92)
- 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)
- TANGO: Co-Speech Gesture Video Reenactment with Hierarchical Audio Motion Embedding and Diffusion Interpolation (ICLR 2025 · similarity 0.90)
- DiffSHEG: A Diffusion-Based Approach for Real-Time Speech-driven Holistic 3D Expression and Gesture Generation (CVPR 2024 · similarity 0.89)
- ConvoFusion: Multi-Modal Conversational Diffusion for Co-Speech Gesture Synthesis (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.
