GB/T 46483-2025Information technology - General technical requirements for customer service virtual digital human (English PDF)

3.1 Virtual digital human

Using technologies such as computer graphics, computer vision, voice interaction, and AIGC (Artificial Intelligence Generated Content), we can create images, sounds, and...

After training on motion models, virtual characters are presented on multimodal output devices using real people or computer-driven methods.

3.2

Virtual digital humans designed for specific industry applications and capable of providing customer service.

3.3

A digital human presented in a two-dimensional plane.

Note. 2D virtual digital humans include different styles such as 2D lifelike and 2D cartoon. 2D lifelike digital human images are usually based on real-life images or AIGC-generated content.

The images and videos of the characters are generated using artificial intelligence technology; 2D cartoon digital human figures are usually drawn based on hand-drawing or image processing software.

Made from [material name].

3.4

A digital human presented in three-dimensional form.

Note. 3D virtual digital humans include different styles such as 3D hyper-realistic, 3D lifelike, and 3D cartoon. 3D virtual digital human images are typically based on 3D modeling software or...

AIGC algorithm creation involves complex modeling, material mapping, and rigging operations, or it can be generated through Neural Radiance Field (NeRF).

3.5 Voice wakeup; voice trigger

A voice interaction system that is in audio stream monitoring mode switches to command word recognition and connection mode after detecting specific features or events.

The process continues with speech recognition and other processing states.

[Source. GB/T 36464.1-2020, 3.17]

3.6

A speech processing technique for analyzing and determining the effective start and end points of speech in a continuous audio stream.

[Source. GB/T 36464.1-2020, 3.22]

3.7 Speech recognition

The process of converting human voice signals into text or instructions.

[Source. GB/T 21023-2007, 3.1]

3.8

Understand the semantic information of data symbols, or the expression of requirements in specific business scenarios, and output correct feedback results as required.

process.

[Source. GB/T 36464.3-2018, 3.6]

3.9 Speech synthesis

The process of synthesizing human language through mechanical and electronic methods.

[Source. GB/T 21024-2007, 3.1]

3.10 gesture recognition

The gesture is determined from the input gesture data.

[Source. GB/T 38665.1-2020, 3.5]

3.11 Limb recognition

The body movements are determined from the input body data.

3.12 Affective interaction

The human-computer interaction process that utilizes user emotions or satisfies user emotional needs.

[Source. GB/T 40691-2021, 3.3]

3.13 modeling

The practice of creating models that represent three-dimensional shapes in a manipulable form.

[Source. GB/T 5271.13-2008, 13.01.08, with modifications]

3.14 Rendering

The process of generating data from 2D/3D models, updating simulations, and rendering the presentation output for a given display device.

[Source. GB/T 44020-2024, 9.1]

3.15 motion capture

A method that tracks and samples target actions in a video sequence and uses this sampling information to represent continuous actions.

[Source. GB/T 41864-2022, 3.6.3.11]