[CVPR 2026] PersonaLive! : Expressive Portrait Image Animation for Live Streaming
update
Zhiyuan Li31417ff4097a
diff --git a/README.md b/README.mdnew file mode 100644index 0000000..9f0f562--- /dev/null+++ b/README.md@@ -0,0 +1,34 @@+<div align="center">++<img src="assets/header.svg" alt="PersonaLive" width="100%">++<h2>Expressive Portrait Image Animation for Live Streaming</h2>+++#### [Zhiyuan Li<sup>1,2,3</sup>](https://huai-chang.github.io/) · [Chi-Man Pun<sup>1</sup>](https://cmpun.github.io/) 📪 · [Chen Fang<sup>2</sup>](http://fangchen.org/) · [Jue Wang<sup>2</sup>](https://scholar.google.com/citations?user=Bt4uDWMAAAAJ&hl=en) · [Xiaodong Cun<sup>3</sup>](https://vinthony.github.io/academic/) 📪+<sup>1</sup> University of Macau <sup>2</sup> Dzine.ai <sup>3</sup> [GVC Lab, Great Bay University](https://gvclab.github.io/)++<img src="assets/demo_1.gif" width="40%"> <img src="assets/demo_2.gif" width="40%">+</div>++## 📣 Updates+- **[2025.12.10]** 🔥 Release `inference code`, `config` and `checkpoints`!++## 🤪 Framework+<img src="assets/overview.png" alt="Image 1" width="100%">+<p align="justify">+We present PersonaLive, a few-step diffusion-based framework for real-time and streamable portrait animation, which achieves low-latency and stable long-term quality.+</p>+++## 🚀 Getting Started++<!-- ## 💖 Citation+If you find PersonaLive useful for your research, welcome to 🌟 this repo and cite our work using the following BibTeX:+```bibtex++``` -->+++## ❤️ Acknowledgement+This project is based on [X-NeMo](https://byteaigc.github.io/X-Portrait2/), [StreamDiffusion](https://github.com/cumulo-autumn/StreamDiffusion) and [RAIN](https://pscgylotti.github.io/pages/RAIN/), thanks to their invaluable contributions.\ No newline at end of filediff --git a/assets/demo_1.gif b/assets/demo_1.gifnew file mode 100644index 0000000..5eb1a8dBinary files /dev/null and b/assets/demo_1.gif differdiff --git a/assets/demo_2.gif b/assets/demo_2.gifnew file mode 100644index 0000000..afdfb40Binary files /dev/null and b/assets/demo_2.gif differdiff --git a/assets/header.svg b/assets/header.svgnew file mode 100644index 0000000..2fb9edb--- /dev/null+++ b/assets/header.svg@@ -0,0 +1,62 @@+<svg fill="none" viewBox="0 0 800 70" width="100%" height="70" xmlns="http://www.w3.org/2000/svg">+ <foreignObject width="100%" height="100%">+ <div xmlns="http://www.w3.org/1999/xhtml">+ <style>+ .rainbow-text-container {+ display: flex;+ justify-content: center;+ align-items: center;+ height: 100%;+ background-color: transparent;+ }++ .rainbow-text {+ font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Helvetica, Arial, sans-serif;+ font-size: 70px;+ font-weight: 900;+ text-align: center;++ /* 1. 颜色列表修改:在最后添加了起始色 #FF6B6B 以实现首尾无缝衔接 */+ background-image: linear-gradient(+ to right,+ #FF6B6B,+ #FF9E42,+ #F0C330,+ #6BCB77,+ #4D96FF,+ #9D78E3,+ #FF6B6B+ );++ /* 2. 背景尺寸设为 200%,为流动提供空间 */+ background-size: 200% auto;++ /* 3. 添加动画:4秒完成一圈,线性匀速,无限循环 */+ animation: rainbow-flow 4s linear infinite;++ -webkit-background-clip: text;+ background-clip: text;+ -webkit-text-fill-color: transparent;+ color: transparent;++ /* 投影保持不变 */+ filter: drop-shadow(0px 2px 4px rgba(0,0,0,0.15));+ }++ /* 4. 定义关键帧动画 */+ @keyframes rainbow-flow {+ 0% {+ background-position: 0% center;+ }+ 100% {+ /* 移动到 -200% 刚好走完一个完整的颜色循环,回到起点 */+ background-position: -200% center;+ }+ }+ </style>+ <div class="rainbow-text-container">+ <div class="rainbow-text">PersonaLive!</div>+ </div>+ </div>+ </foreignObject>+</svg>\ No newline at end of filediff --git a/assets/overview.png b/assets/overview.pngnew file mode 100644index 0000000..747632fBinary files /dev/null and b/assets/overview.png differdiff --git a/demo/demo_1.png b/demo/demo_1.pngnew file mode 100644index 0000000..f5e8cb2Binary files /dev/null and b/demo/demo_1.png differdiff --git a/demo/demo_2.png b/demo/demo_2.pngnew file mode 100644index 0000000..1c01a62Binary files /dev/null and b/demo/demo_2.png differ