Automated Medical Image Analysis Tool Generation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The current manual process of generating machine learning tools for radiology is time-consuming and prone to false negatives, especially when dealing with rare medical findings, as it relies on human intervention and dataset collection for specific purposes.

Innovation Solution

A system that automatically generates multiple data analysis tools based on medical image data and analysis data, using neural networks and clustering algorithms to create tools adapted for specific purposes, reducing human intervention and improving accuracy by analyzing medical image data without manual interaction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual measurement tools are used to collect datasets for training machine learning models, then the models can be trained for specific radiology tasks, but the process becomes time-consuming and expensive

Engineering Contradiction:
Improvemodel training accuracyVSAvoiddata collection time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system enables automatic self-training of machine learning models by utilizing existing clinical workflow data without requiring manual annotation. The tool generation unit automatically generates training datasets from routine clinical measurements, eliminating the need for time-consuming manual data collection while maintaining model training quality

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system performs preliminary data preparation and tool generation in advance by automatically creating training datasets from existing clinical data. This preliminary action prepares the models before actual clinical use, reducing the time required for both data collection and model deployment

Inventive Principle:
Principle #10Preliminary action

2Device complexity

If a single universal detector is used for all clinical measurements, then the system complexity is reduced, but rare findings may be missed causing false negatives

Engineering Contradiction:
Improvesystem structureVSAvoiddetection accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The system segments the detection task by automatically generating multiple specialized data analysis tools, each optimized for specific radiology findings or measurement types. This segmentation allows the system to maintain low overall complexity while achieving high detection accuracy for rare findings through specialized detectors

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically adapts the number and type of detectors based on the specific clinical task and data characteristics. The tool generation unit can create or select appropriate detectors on-demand, allowing the system to balance complexity and accuracy dynamically rather than being fixed

Inventive Principle:
Principle #15Dynamics

3Reliability

If multiple specialized data analysis tools are generated for different medical findings, then detection accuracy improves, but the system complexity and computational resources increase

Engineering Contradiction:
Improvedetection accuracyVSAvoidnumber of tools
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The tool generation unit serves as a universal component that can generate multiple types of specialized detectors through a single automated process. This multi-functional approach allows the system to maintain high detection accuracy across different findings while managing complexity through a unified tool generation mechanism rather than separate development processes

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4339882A1System for medical data analysis
Publication Date: 2024.03.20 SIEMENS HEALTHINEERS AG
  • EP4339882A1 patent drawingFigure 1
  • EP4339882A1 patent drawingFigure 2
  • EP4339882A1 patent drawingFigure 3

AI summary

A system (100) for medical data analysis, comprising a tool generation unit (300) configured for automatically generating a first number of data analysis tools (301, 302) based on first medical image data (IMG1-6) and first analysis data (DL1-6) related to the first medical image data (IMG1-6).