PersonaLive！面向直播的富有表现力肖像图像动画
PersonaLive! Expressive Portrait Image Animation for Live Streaming
面向实时流媒体场景，用少步外观蒸馏压缩去噪步数、用自回归微块流式生成加滑动训练与历史关键帧机制保证长时稳定，这套「蒸馏提速 + 分块流式」组合可直接迁移到其他需要低延迟长视频生成的扩散模型上。
原文依据 a fewer-step appearance distillation strategy is proposed to eliminate appearance redundancy in the denoising process
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论当前基于扩散的肖像动画模型主要关注提升视觉质量与表情真实感，却忽视了生成延迟与实时性能，这限制了其在直播场景中的应用范围。我们提出 PersonaLive，一个面向流式实时肖像动画的新型扩散框架，并配备多阶段训练方案。具体而言，我们首先采用混合隐式信号，即隐式面部表征与三维隐式关键点，以实现富有表现力的图像级运动控制。随后，我们提出一种少步外观蒸馏策略，以消除去噪过程中的外观冗余，从而大幅提升推理效率。最后，我们引入一种自回归微块流式生成范式，并配备滑动训练策略与历史关键帧机制，以实现低延迟且稳定的长时视频生成。大量实验表明，PersonaLive 取得了最优性能，相比先前的基于扩散的肖像动画模型最高可实现 7-22 倍加速。代码将公开。
英文原文摘要 1161 字符
Current diffusion-based portrait animation models predominantly focus on enhancing visual quality and expression realism, while overlooking generation latency and real-time performance, which restricts their application range in the live streaming scenario. We propose PersonaLive, a novel diffusion-based framework towards streaming real-time portrait animation with multi-stage training recipes. Specifically, we first adopt hybrid implicit signals, namely implicit facial representations and 3D implicit keypoints, to achieve expressive image-level motion control. Then, a fewer-step appearance distillation strategy is proposed to eliminate appearance redundancy in the denoising process, greatly improving inference efficiency. Finally, we introduce an autoregressive micro-chunk streaming generation paradigm equipped with a sliding training strategy and a historical keyframe mechanism to enable low-latency and stable long-term video generation. Extensive experiments demonstrate that PersonaLive achieves state-of-the-art performance with up to **7-22**xspeedup over prior diffusion-based portrait animation models. The code will be publicly available.