Research map / Video / Video generation
ActAvatar: Temporally-Aware Precise Action Control for Talking Avatars
Ziqiao Peng, Yi Chen, Yifeng Ma, Guozhen Zhang, Zhiyao Sun, Zixiang Zhou, Youliang Zhang, Zhengguang Zhou, Zhaoxin Fan, Hongyan Liu, Yuan Zhou, Qinglin Lu and 1 more
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 Phase-Aware Cross-Attention (PACA) (ablation) | base global-prompt model · Action Bench | Hit@Segment: 0.725 -> 0.829 | Improved |
| add PACA (ablation) | base global-prompt model · Action Bench | Action Accuracy: 3.91 -> 5.78 | Improved |
| add PACA (ablation) | base global-prompt model · Action Bench | Temporal Correctness: 6.47 -> 7.12 | Improved |
| add PACA (ablation) | base global-prompt model · Action Bench | Sync-C: 6.37 -> 6.39 | No clear effect |
| add progressive audio-visual alignment (ablation) | PACA model · Action Bench | Sync-C: 6.39 -> 6.57 | Improved |
| add progressive audio-visual alignment (ablation) | PACA model · Action Bench | Hit@Segment: 0.829 -> 0.831 | No clear effect |
| add two-stage training (ablation) | PACA + progressive alignment · Action Bench | Sync-C: 6.57 -> 6.89 | Improved |
| add two-stage training (ablation) | PACA + progressive alignment · Action Bench | Hit@Segment: 0.831 -> 0.854 | Improved |
| add two-stage training (ablation) | PACA + progressive alignment · Action Bench | Action Accuracy: 5.75 -> 5.97 | Improved |
| ActAvatar vs best baseline on lip sync | ActAvatar · HDTF test set | Sync-C: 7.663 vs 7.986 (OmniAvatar) | No clear effect |
| ActAvatar vs Wan-S2V on visual quality | ActAvatar · HDTF test set | FID: 23.850 -> 23.471 | Improved |
| ActAvatar with 5B params vs 14B models | ActAvatar · HDTF test set | inference time: 68 min (Wan-S2V) -> 16 min | Improved |
Closest papers by meaning
- CtrlCoMo: Controllable Co-Speech Motion Generation with Gesture–Action Disentanglement (ECCV 2026 · similarity 0.88)
- Making Avatars Interact: Towards Text-Driven Human-Object Interaction for Controllable Talking Avatars (ECCV 2026 · similarity 0.88)
- InstructAvatar: Text-Guided Emotion and Motion Control for Avatar Generation (AAAI 2025 · similarity 0.87)
- SyncDreamer: Controllable and Expressive Avatar Generation Beyond the Talking Head (CVPR 2026 · similarity 0.87)
- Avatar Forcing: Real-Time Interactive Head Avatar Generation for Natural Conversation (CVPR 2026 · similarity 0.87)
- GAIA: Zero-shot Talking Avatar Generation (ICLR 2024 · similarity 0.85)
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.
