Neural Network Failure Detection in MR Scanners
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing failures in medical imaging machines like MR scanners are manual, error-prone, and lack sensitivity and specificity, often requiring expert knowledge and distinguishing between system errors and poor parameterization issues.
Innovation Solution
An automated failure analysis method using a trained neural network that calculates performance indicators from input data, including image and metadata, to identify failure sources with probability measures, allowing for early detection and differentiation between system errors and parameterization issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual failure analysis by service engineers is used, then expert knowledge can be applied to diagnose system errors, but the process is error-prone and relies heavily on technician experience
Solution Approach 1:
The system enables automatic self-diagnosis of MR scanner failures through a neural network that autonomously analyzes performance indicators and identifies failure sources without requiring manual intervention by service engineers, thereby improving reliability while reducing operational complexity
Solution Approach 2:
The patent replaces the manual mechanical process of service engineer diagnosis with an automated computational system using neural networks and performance indicator analysis, substituting human expertise with an automated intelligent system that consistently identifies failures without human error
2Reliability
If automated failure analysis is implemented, then manual errors are reduced and consistency is improved, but the system complexity increases
Solution Approach 1:
The neural network system serves multiple functions simultaneously: it monitors system performance, detects failures, identifies failure sources, and provides diagnostic recommendations, thereby achieving high reliability through a single multi-functional automated system rather than multiple separate systems
Solution Approach 2:
The patent introduces performance indicators as intermediary metrics that bridge the gap between raw system data and failure diagnosis, allowing the neural network to automatically translate complex system states into actionable diagnostic information without requiring direct complex analysis of all system parameters
3Loss of time
If performance indicators are calculated from raw data before image reconstruction, then early detection is achieved and costs are reduced, but the processing time requirements increase
Solution Approach 1:
The system performs preliminary calculation of performance indicators from raw data immediately after acquisition, before image reconstruction occurs. This early detection capability allows failures to be identified at the earliest possible stage, reducing downtime and costs while the neural network processes the extracted indicators to identify potential issues
Data Source
AI summary
In a method, a computer and a medical computer for automatic failure analysis in order to provide a cause of failure of the medical imaging apparatus during operation, input data are read into the computer that include raw data or image data, acquired by the imaging apparatus. A set of performance indicators in the input data is calculated by the computer. A trained neural network system is accessed with the calculated performance indicators, in order to provide result data that, in the case of a failure, identify a failure source.