Connecting the world and beyond

Digital Human Systems

7.0.1.jpg Focuses on the standardized construction and service systems for virtual/digital humans. It systematically standardizes comprehensive requirements from 2D/3D digital human asset modeling, multimodal AI driving of natural language and actions, tactile feedback rendering engines, to unified access interface standards across terminal platforms.

7.1 Digital Human Platforms

Formulates the general system reference architecture for digital humans, as well as standards for rendering precision grading, system concurrency, and O&M management of cloud-based digital human SaaS service platforms driven by foundation models.

ITU-T F.748.15: Framework and metrics for digital human application systems (2022)
Proposes a general technical module breakdown (including image, speech, and animation modules) and subjective/objective testing evaluation metrics for digital human application systems.

ITU-T F.748.63: Framework and general technical requirements of foundation model enhanced digital human system (2025)
Standardizes the architectural requirements for utilizing foundation model capabilities to enhance digital human systems, encompassing everything from multimodal intent recognition to the natural generation of fluent gestures and facial expressions.

F.748.63 Framework of foundation model enhanced digital human systems

ITU-T F.748.64: Requirements of digital human service platform (2025)
Targets digital human Software-as-a-Service (SaaS) cloud platforms, providing detailed grading standards for application deployment reliability, model customization functions, and 3D digital human modeling/rendering quality (e.g., head models, hair).

7.2 Access Interfaces & Interaction Evaluation

Standardizes the system testing methods for non-interactive virtual humans and smartphone-based 3D digital humans, along with audio/video interaction collaboration architectures in IoT communications and tactile immersive rendering system constraints.

ITU-T F.748.14: Requirements and evaluation methods of non-interactive 2D real-person digital human application systems (2022)
Specifies technical testing and rating formulas for the action fluency, lip-sync audio/video synchronization, and character fidelity of non-interactive (e.g., broadcasting) 2D real-person digital humans.

ITU-T F.748.31: Technical requirements and evaluation methods of 3D digital human system based on smart mobile devices (2024)
Provides functional architecture and model volume evaluation specifications for real-time rendering and interactive driving of 3D digital human systems on mobile device terminals with limited computing power.

ITU-T F.748.50: Framework and requirements of digital human access interfaces (2025)
To achieve flexible distribution and integration of digital human platform capabilities, standardizes the network stream access interface protocol framework connecting the application layer and underlying infrastructure.

F.748.50 Digital human access interface framework

ITU-T F.748.42: Requirements and framework for virtual tactile interaction systems (2025)
Standardizes the system functional architecture constraints for realizing virtual force/tactile feedback interaction loops in immersive virtual multimedia system environments.

ITU-T F.748.51: Requirements for artificial intelligence-based tactile rendering system in multimedia terminals (2025)
Establishes a reference framework for integrating AI prediction model-based tactile rendering services in multimedia smart terminals.

ITU-T F.748.6: Requirements and framework for interactive multimedia communication system of Internet of things (IoT) devices (2025)
Targeting collaboration scenarios among massive IoT devices, establishes multimedia communication platform signaling and architecture specifications integrating audio/video data transmission and AI interaction (e.g., voice recognition interaction).

ITU-T F.AI-SCVC: Requirements and framework of AI-based semantic communication for video calls (2025)
Standardizes a reference technical framework that uses AI algorithms to extract key semantic features of video frames for video call reconstruction under ultra-low bandwidth conditions.