---
license: apache-2.0
tags:
- Portrait Animation
- Video Generation
---

# LIA-X: Interpretable Latent Portrait Animator
[Yaohui Wang](https://wyhsirius.github.io/), [Di Yang](https://walker1126.github.io/), [Xinyuan Chen](https://scholar.google.com/citations?user=3fWSC8YAAAAJ&hl=zh-CN), [François Brémond](https://www-sop.inria.fr/members/Francois.Bremond/), [Yu Qiao](https://scholar.google.com/citations?user=gFtI-8QAAAAJ&hl=en), [Antitza Dantcheva](https://www-sop.inria.fr/members/Antitza.Dantcheva/)

[![arXiv](https://img.shields.io/badge/arXiv-2508.09959-b31b1b.svg)](https://arxiv.org/abs/2508.09959)
[![Project Page](https://img.shields.io/badge/Project-Website-green)](https://wyhsirius.github.io/LIA-X-project/)
[![Hugging Face Spaces](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-yellow)](https://huggingface.co/spaces/YaohuiW/LIA-X)

LIA-X is a novel latent portrait animator, which supports interpretable and fine-grained control of **head**, **mouth** and **eyes**.

<img src="teaser.gif" width="1000">