Saved in:
Bibliographic Details
Main Authors: Chen, Peiyin, Yang, Zhuowei, Feng, Hui, Jiang, Sheng, Yan, Rui
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.10650
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912644171038720
author Chen, Peiyin
Yang, Zhuowei
Feng, Hui
Jiang, Sheng
Yan, Rui
author_facet Chen, Peiyin
Yang, Zhuowei
Feng, Hui
Jiang, Sheng
Yan, Rui
contents Audio-driven talking-head generation has advanced rapidly with diffusion-based generative models, yet producing temporally coherent videos with fine-grained motion control remains challenging. We propose DEMO, a flow-matching generative framework for audio-driven talking-portrait video synthesis that delivers disentangled, high-fidelity control of lip motion, head pose, and eye gaze. The core contribution is a motion auto-encoder that builds a structured latent space in which motion factors are independently represented and approximately orthogonalized. On this disentangled motion space, we apply optimal-transport-based flow matching with a transformer predictor to generate temporally smooth motion trajectories conditioned on audio. Extensive experiments across multiple benchmarks show that DEMO outperforms prior methods in video realism, lip-audio synchronization, and motion fidelity. These results demonstrate that combining fine-grained motion disentanglement with flow-based generative modeling provides a powerful new paradigm for controllable talking-head video synthesis.
format Preprint
id arxiv_https___arxiv_org_abs_2510_10650
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DEMO: Disentangled Motion Latent Flow Matching for Fine-Grained Controllable Talking Portrait Synthesis
Chen, Peiyin
Yang, Zhuowei
Feng, Hui
Jiang, Sheng
Yan, Rui
Computer Vision and Pattern Recognition
Artificial Intelligence
Audio-driven talking-head generation has advanced rapidly with diffusion-based generative models, yet producing temporally coherent videos with fine-grained motion control remains challenging. We propose DEMO, a flow-matching generative framework for audio-driven talking-portrait video synthesis that delivers disentangled, high-fidelity control of lip motion, head pose, and eye gaze. The core contribution is a motion auto-encoder that builds a structured latent space in which motion factors are independently represented and approximately orthogonalized. On this disentangled motion space, we apply optimal-transport-based flow matching with a transformer predictor to generate temporally smooth motion trajectories conditioned on audio. Extensive experiments across multiple benchmarks show that DEMO outperforms prior methods in video realism, lip-audio synchronization, and motion fidelity. These results demonstrate that combining fine-grained motion disentanglement with flow-based generative modeling provides a powerful new paradigm for controllable talking-head video synthesis.
title DEMO: Disentangled Motion Latent Flow Matching for Fine-Grained Controllable Talking Portrait Synthesis
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2510.10650