Neural ARDS Detection Using Thoracic Images and Vital Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting acute respiratory distress syndrome (ARDS) in intensive care patients are unreliable, as they rely solely on X-ray images or physiological parameters, failing to provide early and accurate detection.

Innovation Solution

A computer system utilizing an artificial neural network with multiple subnetworks processes thoracic radiological images and vital data over time to generate an ARDS indicator value, which is compared to a threshold for early detection and notification of potential ARDS.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If automated detection systems use only X-ray images or only physiological parameters, then the system complexity is reduced, but the detection reliability is insufficient

Engineering Contradiction:
ImproveARDS detection reliabilityVSAvoiddetection system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines multiple data sources (X-ray images, physiological parameters, laboratory values) into a unified deep learning model that processes all inputs simultaneously. The system merges radiological image analysis with physiological parameter monitoring to achieve reliable ARDS detection, resolving the contradiction by integrating rather than selecting单一 data source.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The detection system is segmented into multiple specialized subnetworks: a first subnetwork for processing radiological images, a second subnetwork for physiological parameters, and a third subnetwork for laboratory values. Each subnetwork processes specific data types independently before merging results, allowing complex multi-source analysis while maintaining modular architecture.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If deep learning models process multiple data sources simultaneously, then detection accuracy improves, but computational resource requirements increase

Engineering Contradiction:
ImproveARDS detection accuracyVSAvoidcomputational energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The computational workload is segmented across multiple specialized subnetworks, each optimized for specific data types. This segmentation allows parallel processing of different data sources without requiring a single monolithic model, reducing peak computational demands while maintaining high detection accuracy through comprehensive multi-source analysis.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4070327B1Assistance with detection of lung diseases
Publication Date: 2025.12.17 BAYER AG
  • EP4070327B1 patent drawingFigure 1~2
  • EP4070327B1 patent drawingFigure 3~4
  • EP4070327B1 patent drawingFigure 5~6

AI summary

The invention relates to the detection of an acute respiratory distress syndrome in a patient. The subject matter of the present invention relates to a computer system, a method and a computer program product for detecting an acute respiratory distress syndrome.