Xemo-Talker system unlocks explicit emotion control in audio-driven talking portraits

Researchers have developed Xemo-Talker, a novel system for generating audio-driven talking portraits with explicit emotion control. The system addresses the challenge of balancing accurate lip synchronization with fine-grained emotional expression by focusing supervision on less-principal components of motion, which encode emotional cues rather than primary articulation. Xemo-Talker achieves state-of-the-art emotion classification accuracy while maintaining competitive lip synchronization and high inference efficiency, with its source code available on Hugging Face. AI

IMPACT This research could lead to more expressive and controllable AI-generated avatars for virtual communication and entertainment.

RANK_REASON The cluster contains a research paper detailing a new model for audio-driven talking portrait synthesis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• Hugging Face
• Xemo-Talker

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →