Research map / Video / Video generation
StreamAvatar: Streaming Diffusion Models for Real-Time Interactive Human Avatars
Zhiyao Sun, Ziqiao Peng, Yifeng Ma, Yi Chen, Zhengguang Zhou, Zixiang Zhou, Guozhen Zhang, Youliang Zhang, Yuan Zhou, Qinglin Lu, Yong-Jin Liu
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 Sink to autoregressive student (ablation) | StreamAvatar · Talking avatar long set (25 imgs, 20s) | IQA: 4.01 -> 4.64 | Improved |
| add Reference Sink (ablation) | StreamAvatar · Talking avatar short set (50 imgs, 5s) | FID: 96.58 -> 88.75 | Improved |
| add RAPR positional re-encoding (ablation) | StreamAvatar · Talking avatar short set (50 imgs, 5s) | FID: 88.75 -> 81.63 | Improved |
| add RAPR positional re-encoding (ablation) | StreamAvatar · Talking avatar long set (25 imgs, 20s) | IQA: 4.64 -> 4.84 | Improved |
| add adversarial refinement with standard discriminator (ablation) | StreamAvatar · Talking avatar short set (50 imgs, 5s) | FID: 81.63 -> 79.68 | Improved |
| add adversarial refinement with standard discriminator (ablation) | StreamAvatar · Talking avatar long set (25 imgs, 20s) | HA: 0.961 -> 0.993 | Mixed |
| use consistency-aware discriminator instead of standard discriminator (ablation) | StreamAvatar · Talking avatar short set (50 imgs, 5s) | FID: 79.68 -> 74.21 | Improved |
| use consistency-aware discriminator instead of standard discriminator (ablation) | StreamAvatar · Talking avatar long set (25 imgs, 20s) | Sync-C: 6.38 -> 6.64 | Improved |
| apply audio mask to Wav2Vec features instead of raw audio (ablation) | StreamAvatar (undistilled) · Interactive avatar audio mask position | LHKV: 15.49 -> 21.44 | Improved |
| add listening-audio modules vs silent-audio baseline | OmniAvatar baseline · Interactive avatar (50 SpeakerVid-5M videos) | LBKV: 6.05 -> 15.88 | Improved |
| 3-step causal distilled model vs slower offline baselines | StreamAvatar · Talking avatar short set (50 imgs, 5s) | FVD: 557.46 -> 707.34 | Hurt |
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.
- Hallo3 — Hallo3: Highly Dynamic and Realistic Portrait Image Animation with Video Diffusion Transformer (CVPR 2025) · better on 1 of 1 shared result
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
- VLOGGER: Multimodal Diffusion for Embodied Avatar Synthesis (CVPR 2025 · similarity 0.88)
- DAWN: Dynamic Frame Avatar with Non-autoregressive Diffusion Framework for Talking head Video Generation (ICLR 2025 · similarity 0.88)
- Teller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation (CVPR 2025 · similarity 0.88)
- FADA: Fast Diffusion Avatar Synthesis with Mixed-Supervised Multi-CFG Distillation (CVPR 2025 · similarity 0.87)
- RAM-Avatar: Real-time Photo-Realistic Avatar from Monocular Videos with Full-body Control (CVPR 2024 · similarity 0.87)
- AvatarGO: Zero-shot 4D Human-Object Interaction Generation and Animation (ICLR 2025 · similarity 0.87)
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.
