YY/T 1833.1-2022Artificial intelligence medical device—Quality requirements and evaluation - Part 1: Terminology (English PDF)

3.1 Basic technical terms

3.1.1 artificial intelligence; AI

ability of functional units that exhibit various functions related to human intelligence (such as inference and learning)

[Source: GB/T 5271.28-2001, 28.01.02]

3.1.2 artificial intelligence medical device; AIMD

medical device that uses AI technology to realize their intended use

Note 1: For example, independent software that uses machine learning, pattern recognition, rule inference and other technologies to realize its medical use.

Note 2: Medical devices that use embedded AI algorithm and AI chip to realize their medical use.

3.1.3 medical device software

developed software system integrated into a medical device under development, or software system intended to be used as a medical device

Note: Medical device software includes software components and independent software, where software components refer to software embedded in a medical device or used as a part of the medical device, and independent software refers to software that has one or more medical purposes, can complete its expected purpose without medical device hardware, and runs on a general computing platform.

[Source: YY/T 0664-2020, 3.11]

3.1.4 pattern recognition

recognition of physical or abstract patterns, structure and configuration of an object through functional units

[Source: GB/T 5271.28-2001, 28.01.13]

3.1.5 artificial neural network; ANN

network of basic processing elements connected by weighted links with adjustable weights, which generates a value for each unit by applying a non-linear function to its input value and transmits it to other units or expresses it as an output value

Note: Also referred to as neural network.

[Source: GB/T 5271.34-2006, 34.01.06]

3.1.6 inference

method for deriving a conclusion from a known premise

Note 1: In the field of AI, the premise is a fact or a rule.

Note 2: The term "inference" refers to both process and result.

[Source: GB/T 5271.28-2001, 28.03.01, modified]

3.1.7 feature

measurable attribute that can express the essential functional or structural characteristics of a pattern

Note: Such as size, texture, shape, etc. A good feature can cluster similar patterns and separate different patterns.