Research map / Video / Video generation
StreamingTalker: Audio-driven 3D Facial Animation with Autoregressive Diffusion Model
Yifan Yang, Zhi Cen, Sida Peng, Xiangwei Chen, Yifu Deng, Xinyu Zhu, Fan Jia, Xiaowei Zhou, Hujun Bao
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 diffusion head instead of direct MLP prediction (ablation) | StreamingTalker · BIWI | FDD: 5.3934 -> 3.6690 | Improved |
| add VQ-VAE latent space instead of raw motion denoising (ablation) | StreamingTalker · BIWI | LVE: 8.4578 -> 4.2504 | Improved |
| use VQ-VAE instead of standard VAE (ablation) | StreamingTalker · BIWI | LVE: 4.2686 -> 4.2504 | No clear effect |
| use fixed-length history instead of all history (ablation) | StreamingTalker · BIWI | LVE: 4.7598 -> 4.2504 | Improved |
| keep cross-attention between audio and motion (ablation) | StreamingTalker · BIWI | LVE: 11.7510 -> 4.2504 | Improved |
| keep self-attention layers (ablation) | StreamingTalker · BIWI | MOD: 15.5354 -> 8.5208 | Improved |
| keep self-attention layers (ablation) | StreamingTalker · BIWI | FDD: 4.1083 -> 3.6690 | Improved |
| autoregressive diffusion with fixed history instead of full-sequence diffusion | StreamingTalker vs DiffSpeaker/FaceDiffuser · BIWI-Test-B long sequences (2000 frames) | LVE: 5.2213 -> 4.4596 (vs DiffSpeaker) | Improved |
| autoregressive streaming generation | StreamingTalker · Inference latency, 3-27 s audio | latency: 25 ms first frame, constant across audio length | Improved |
| StreamingTalker vs DiffSpeaker on short sequences | StreamingTalker · BIWI | MOD: 8.4091 -> 8.5208 | Hurt |
Closest papers by meaning
- Teller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation (CVPR 2025 · similarity 0.91)
- KeyFace: Expressive Audio-Driven Facial Animation for Long Sequences via KeyFrame Interpolation (CVPR 2025 · similarity 0.91)
- FaceTalk: Audio-Driven Motion Diffusion for Neural Parametric Head Models (CVPR 2024 · similarity 0.91)
- Probabilistic Speech-Driven 3D Facial Motion Synthesis: New Benchmarks Methods and Applications (CVPR 2024 · similarity 0.90)
- DAWN: Dynamic Frame Avatar with Non-autoregressive Diffusion Framework for Talking head Video Generation (ICLR 2025 · similarity 0.90)
- Diffusion-based Realistic Listening Head Generation via Hybrid Motion Modeling (CVPR 2025 · similarity 0.90)
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.
