Large Language Model-Driven 3D Hyper-Realistic Interactive Intelligent Digital Human System.

Area of Science:

• Digital Transformation
• Artificial Intelligence
• Human-Computer Interaction

Background:

• Digital technologies are integrating across all aspects of society, economy, and culture.
• Artificial intelligence generated content (AIGC) is a key driver of change in digital content production and consumption.
• The convergence of digital and physical worlds necessitates new models for interaction.

Purpose of the Study:

• To present an award-winning large language model (LLM)-powered 3D hyper-realistic interactive digital human system.
• To explore the integration of automatic speech recognition (ASR), natural language processing (NLP), and emotional text-to-speech (TTS) technologies.
• To discuss the application of computer graphics (CG) and artificial intelligence (AI) in creating photorealistic 3D environments for meta humans.

Main Methods:

• Development of a modular, client-server distributed architecture for scalable progress.
• Integration of LLM, ASR, NLP, and emotional TTS for interactive capabilities.
• Utilizing CG and AI to create photorealistic 3D environments and meta humans.

Main Results:

• A functional 3D hyper-realistic interactive digital human system powered by an LLM.
• Demonstration of seamless integration of speech and language technologies.
• Creation of photorealistic 3D environments and meta humans.

Conclusions:

• AIGC and advanced digital human systems are crucial for enhancing digital lives.
• The developed system represents a significant step towards a new era of digital civilization.
• Potential applications span various domains, highlighting the technology's transformative impact.