Research map / Video / Video generation
PersonaLive! Expressive Portrait Image Animation for Live Streaming
Zhiyuan Li, Chi-Man Pun, Chen Fang, Jue Wang, Xiaodong Cun
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 sliding training strategy (ablation) | PersonaLive · LV100 cross-reenactment | ID-SIM: 0.549 -> 0.698 | Improved |
| add historical keyframe mechanism (ablation) | PersonaLive · LV100 cross-reenactment | tLP: 13.27 -> 12.83 tLP; ID-SIM 0.728 -> 0.698 | Mixed |
| add motion-interpolated initialization (ablation) | PersonaLive · LV100 cross-reenactment | tLP: 13.06 -> 12.83 | Improved |
| reduce chunk size from 4 to 2 (ablation) | PersonaLive · LV100 cross-reenactment | ID-SIM: 0.698 -> 0.660 | Hurt |
| replace bidirectional attention with chunk-wise causal attention (ablation) | PersonaLive · LV100 cross-reenactment | ID-SIM: 0.698 -> 0.689 | Hurt |
| distill to few steps without adversarial loss (ablation) | PersonaLive appearance distillation · qualitative (Fig. 6) | visual quality | Hurt |
| add adversarial loss to distillation (ablation) | PersonaLive appearance distillation · qualitative (Fig. 6) | visual fidelity | Improved |
| apply CFG at inference instead of adversarial loss (ablation) | PersonaLive · Efficiency (single H100) | FPS: 15.82 -> 9.5 | Hurt |
| reduce denoising steps without distillation (ablation) | PersonaLive · qualitative (Fig. 6) | visual quality | Hurt |
| replace VAE decoder with TinyVAE | PersonaLive · Efficiency (single H100) | FPS: 15.82 -> 20 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- HunyuanPortrait — HunyuanPortrait: Implicit Condition Control for Enhanced Portrait Animation (CVPR 2025) · better on 3 of 3 shared results
- X-NeMo — X-NeMo: Expressive Neural Motion Reenactment via Disentangled Latent Attention (ICLR 2025) · better on 0 of 2 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
- RealPortrait: Realistic Portrait Animation with Diffusion Transformers (AAAI 2025 · similarity 0.90)
- Teller: Real-Time Streaming Audio-Driven Portrait Animation with Autoregressive Motion Generation (CVPR 2025 · similarity 0.88)
- Hallo3: Highly Dynamic and Realistic Portrait Image Animation with Video Diffusion Transformer (CVPR 2025 · similarity 0.88)
- HunyuanPortrait: Implicit Condition Control for Enhanced Portrait Animation (CVPR 2025 · similarity 0.88)
- AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning (ICLR 2024 · similarity 0.86)
- High-Fidelity Relightable Monocular Portrait Animation with Lighting-Controllable Video Diffusion Model (CVPR 2025 · 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.
