REA-Listener: Real-Time Listening Head Generation with Dynamic Emotion Modeling and Flexible Modality Adaptation

Open-Source TODO List

• 🟩 Model code
• 🟩 Inference code
• ⬜️ Environment setup
• ⬜️ Checkpoints
• ⬜️ Training guide for custom data

Acknowledgments

This repository was developed with reference to the following projects, We thank the original authors for their valuable work: ViCo Challenge Baseline, ListenFormer, GAGAvatar.

📚Citation

@inproceedings{realistener2025,
  title={REA-Listener: Real-Time Listening Head Generation with Dynamic Emotion Modeling and Flexible Modality Adaptation},
  author={Zhao, Sizhe and Wang, Chenyang and Zhao, Weiyu and Li, Zonglin and Li, Ming and Zhang, Shengping},
  booktitle={Proceedings of the 33rd ACM International Conference on Multimedia},
  pages={9733--9742},
  year={2025}
}