Brain MRI Neural Network Prognosis Without Manual ROI Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing brain MRI analysis methods for cardiac arrest prognosis face challenges due to low practicality and difficulty in obtaining training data, particularly in clinical settings where cardiac arrest patient recovery rates are low, and require complex user intervention for ROI selection.

Innovation Solution

A neural network-based brain MRI analysis device and method that utilizes a convolutional neural network (CNN) to analyze DWI, ADC, and FA images, generating input data through concatenation of 3D arrays, and employs multiple neural network models to determine neurological prognosis without human intervention, using both cardiac arrest and non-cardiac arrest patient data for training.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual ROI selection by user is used, then measurement precision is improved, but device complexity and ease of operation deteriorate

Engineering Contradiction:

Improveprognosis determination accuracyVSAvoidease of operation

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The neural network model automatically performs ROI selection and prognosis determination without requiring manual user intervention. The system processes brain MRI images autonomously, extracting relevant features and generating prognostic assessments independently, thereby eliminating the need for manual ROI marking while maintaining diagnostic accuracy.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of user-driven ROI selection with an automated neural network-based system. The convolutional neural network automatically identifies and processes regions of interest in brain MRI images, substituting human operator intervention with an automated computational system that achieves consistent and reproducible results.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If dedicated software or algorithms for ROI selection are used, then measurement precision is improved, but device complexity worsens

Engineering Contradiction:

Improveprognosis determination accuracyVSAvoiddevice complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The neural network model serves multiple functions within a single integrated system: it performs automatic ROI identification, feature extraction, and prognosis determination simultaneously. This multi-functional approach eliminates the need for separate dedicated software modules for ROI selection and analysis, thereby reducing overall system complexity while maintaining high measurement precision.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent merges the previously separate tasks of ROI selection and prognosis analysis into a unified neural network-based system. By combining these functions into a single integrated model that processes brain MRI images end-to-end, the system reduces the complexity of having multiple dedicated algorithms while achieving consistent prognostic accuracy.

Inventive Principle:

Principle #5Merging (Combining)

3Reliability

If training data from cardiac arrest patients is used, then reliability is improved, but quantity of substance deteriorates

Engineering Contradiction:

Improvemodel training reliabilityVSAvoidtraining data quantity

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The patent introduces non-cardiac arrest patient data as an intermediary resource to supplement and expand the training dataset. By incorporating data from patients without cardiac arrest, the system increases the overall quantity of training data available while maintaining the inclusion of cardiac arrest patient data for reliable prognostic modeling. This intermediary data source enables more comprehensive training without compromising the specificity of cardiac arrest prognosis.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20260033787A1Brain MRI analysis method and device using neural network

Publication Date: 2026.02.05 SEOUL NAT UNIV HOSPITAL

AI summary

A brain magnetic resonance imaging (MRI) analysis device and method using a neural network are disclosed. The brain MRI analysis device using a neural network according to one embodiment comprises: a memory for storing a neural network model; and a processor, which is connected to the memory so as to control an analysis device, wherein the processor receives one or more brain MRI images so as to generate input data, and inputs the input data into the neural network model so as to acquire output data, and the neural network model is trained to determine neurological prognosis of a cardiac arrest patient if the input data is input into the neural network model.