Medical Image Metadata Standardization via Trained Function

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Inconsistent and incorrect metadata attributes in medical image data hinder cross-regional sharing and analysis, as they are often filled in locally without standardization, leading to difficulties in comparing and interpreting medical images across different healthcare entities.

Innovation Solution

An automated machine learning pipeline that maps attribute values and tags to standard names or codes, utilizing a trained function to translate local medical terms into a common standard, ensuring unified and standardized metadata attributes for medical image data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If metadata attributes are filled in manually by operators using local languages and knowledge, then the filling process is simple and quick, but the attribute values become inconsistent and non-standardized across different regions

Engineering Contradiction:

Improvemetadata filling efficiencyVSAvoidmetadata standardization

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

A trained function acts as an intermediary between the provisional attribute value (local language input) and the standardized attribute value. This function automatically translates and standardizes the metadata without requiring manual intervention, thus maintaining filling efficiency while ensuring standardization reliability.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs self-service by automatically determining standardized attribute values through a trained function without human intervention. The metadata attribute is self-corrected and self-standardized, eliminating the trade-off between manual filling speed and standardization quality.

Inventive Principle:

Principle #25Self-service

2Reliability

If automated protocols are used to fill metadata attributes, then standardization is improved, but errors occur when protocols are incorrectly applied to different examination objects

Engineering Contradiction:

Improvemetadata standardizationVSAvoidattribute value accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The system uses the medical image data as feedback to verify and correct the provisional attribute value. The trained function analyzes both the protocol-based provisional value and the actual image content, adjusting the final standardized value to ensure accuracy even when protocols are misapplied.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The manual verification process is replaced by an automated machine learning-based trained function that analyzes the medical image data to verify attribute accuracy. This substitution of mechanical/manual verification with intelligent automated analysis improves both standardization and precision.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Adaptability or versatility

If manual filling of metadata attributes is performed, then flexibility in using local expressions is maintained, but querying and comparing medical images across different regions becomes impossible

Engineering Contradiction:

Improvelocal language flexibilityVSAvoidimage query and comparison capability

Core Design Contradiction:

Adaptability or versatilityVSDifficulty of detecting and measuring

Solution Approach 1:

The trained function serves as an intermediary that translates diverse local expressions and provisional attribute values into a unified standardized format. This enables the system to maintain adaptability to local languages during input while ensuring uniformity for querying and comparison operations.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The standardized attribute value system provides universality, allowing the same standard to be applied across all regions and languages. This universal standard enables cross-regional querying and comparison while the system simultaneously handles multiple local languages through the trained function's translation capability.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20240161908A1Method for providing at least one first metadata attribute comprised by medical image data

Publication Date: 2024.05.16 SIEMENS HEALTHINEERS AG

AI summary

One or more example embodiments of the present invention describes a computer-implemented method for providing at least one first metadata attribute associated with medical image data. The method comprises receiving the medical image data and the at least one first metadata attribute, the at least one first metadata attribute including an attribute tag and a provisional attribute value; applying a first trained function to the medical image data to determine a standardized attribute value; determining a final attribute value based on the provisional attribute value and the standardized attribute value; and providing the at least one first metadata attribute, the at least one first metadata attribute including the attribute tag and the final attribute value.