Attention Module Machine Learning Model for Fetal Gestational Age Estimation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In low- and middle-income countries, estimating gestational age is challenging due to the lack of access to traditional ultrasound machines, trained sonographers, and expert interpretation, necessitating a non-expert method for accurately estimating gestational age using ultrasound image data.

Innovation Solution

A trained machine learning model with an attention module is used to receive and process ultrasound image data, producing a weighted sum vector that aggregates and weights feature vectors to estimate gestational age, providing a reliable output to users.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional ultrasound machines and expert sonographers are used to estimate gestational age, then measurement precision is improved, but device complexity and cost increase

Engineering Contradiction:

Improvegestational age estimation accuracyVSAvoidultrasound system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces the mechanical and manual system of traditional ultrasound measurement with an automated machine learning system. The ML model processes ultrasound images and automatically estimates gestational age without requiring expert sonographers to manually measure fetal structures, thereby maintaining measurement precision while reducing device complexity and operational requirements

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service by allowing non-expert users to obtain gestational age estimates through automated processing. The machine learning model performs the complex analysis task independently without requiring skilled operators, making the technology accessible in resource-limited settings where trained sonographers are unavailable

Inventive Principle:

Principle #25Self-service

2Measurement precision

If traditional ultrasound machines and trained sonographers are deployed, then gestational age estimation accuracy is improved, but ease of operation deteriorates due to requiring expert training

Engineering Contradiction:

Improvegestational age estimation accuracyVSAvoidultrasound operation difficulty

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The patent substitutes the need for expert human operation with an automated machine learning system. The ML model handles the complex task of analyzing fetal biometry and calculating gestational age, replacing the need for trained sonographers and making the system easy to operate for non-experts

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The machine learning model acts as an intermediary between the raw ultrasound images and the gestational age estimation. It bridges the gap by automatically performing the complex analysis that previously required expert human interpretation, thereby simplifying the operational process while maintaining accuracy

Inventive Principle:

Principle #24Intermediary (Mediator)

3Ease of operation

If self-reported last menstrual period is used to determine gestational age, then ease of operation is improved, but measurement precision deteriorates due to reliability issues

Engineering Contradiction:

Improvegestational age determination simplicityVSAvoidgestational age accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent replaces the unreliable self-reported LMP method with an automated machine learning analysis of ultrasound images. The ML system objectively measures fetal structures and calculates gestational age, providing accurate results without requiring women to recall or report menstrual dates, thus improving precision while maintaining ease of use

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20250090136A1Methods, systems, and computer readable media for using trained machine learning model including an attention module to estimate gestational age from ultrasound image data

Publication Date: 2025.03.20 THE UNIV OF NORTH CAROLINA AT CHAPEL HILL

AI summary

A method for estimating gestational age of a human fetus using a trained machine learning model with an attention function includes receiving, at a feature extraction module of a trained machine learning model, fetal ultrasound image data for at least one image of a human fetus, and producing, by propagating the ultrasound image data through the feature extraction module, at least one feature vector from the ultrasound image data. The method further includes providing the at least one feature vector as input to an attention module of the trained machine learning model and producing, by propagating the feature vectors through the attention module, a weighted sum vector that aggregates and weights the feature vectors. The method further includes providing the weighted sum vector as input to a gestational age prediction module of the trained machine learning model, which generates, from the weighted sum vector, an estimate of the gestational age of the human fetus. The method further includes outputting the estimate of gestational age to a user.