Neural Network Cardiac Function Quantification from CMR Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining cardiac function from cardiac magnetic resonance (CMR) images require manual annotations and explicit segmentation, which are time-consuming and prone to errors, and do not allow for real-time EF quantification without interrupting ongoing image acquisition.
Innovation Solution
A neural network system that processes temporal sequences of CMR images to output ejection fraction (EF) values directly, trained on datasets including annotated sequences, enabling real-time EF determination without explicit segmentation and allowing for adjustments to the CMR image acquisition protocol during scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual annotations and explicit segmentation are used to determine cardiac function from CMR images, then measurement precision is improved, but loss of time increases and productivity decreases
Solution Approach 1:
The system performs self-service by automatically determining ejection fraction from CMR images without requiring manual annotations. The neural network processes the images directly to extract EF values, eliminating the need for human operators to perform time-consuming segmentation and annotation tasks while maintaining measurement precision.
Solution Approach 2:
The patent replaces the mechanical manual annotation process with an automated computational system. Instead of human operators manually segmenting and annotating cardiac structures, a neural network-based system automatically processes CMR images to determine EF, substituting human mechanical work with automated image processing algorithms.
2Measurement precision
If manual annotations are performed for EF quantification, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system extracts only the essential information needed for EF determination from CMR images without requiring full manual segmentation of cardiac structures. By focusing on extracting key features directly relevant to ejection fraction calculation, the system reduces the complexity of the overall process while maintaining measurement precision.
Solution Approach 2:
The neural network system performs multiple functions simultaneously: it processes CMR images, identifies cardiac structures, calculates volumes, and determines EF values in a single integrated operation. This multi-functionality eliminates the need for separate segmentation and annotation steps, reducing device complexity while maintaining precision.
3Measurement precision
If conventional methods are used for EF quantification, then measurement precision is improved, but adaptability decreases as ongoing image acquisition cannot be interrupted
Solution Approach 1:
The system provides real-time feedback by determining EF values during ongoing CMR image acquisition without interrupting the scan. The automated processing allows the system to analyze images as they are acquired, provide EF measurements in real-time, and enable operators to adjust imaging protocols based on these feedback results while the scan continues.
Solution Approach 2:
The system performs preliminary EF determination during the image acquisition process itself, before the complete scan is finished. By processing images as they are acquired and providing intermediate EF values, the system enables adaptive protocol adjustments without requiring completion of the entire scan sequence or interrupting ongoing acquisition.
Data Source
AI summary
A value indicative of an ejection fraction, EF, of a cardiac chamber of a heart is based on a temporal sequence of cardiac magnetic resonance, CMR, images of the cardiac chamber. A neural network system has an input layer configured to receive the temporal sequence of a stack of slices of the CMR images along an axis of the heart. The temporal sequence is one or multiple consecutive cardiac cycles of the heart. The neural network system has an output layer configured to output the value indicative of the EF based on the temporal sequence. The neural network system has interconnections between the input layer and the output layer and is trained with a plurality of datasets. Each of the datasets comprises an instance temporal sequence of the stack of slices of the CMR images along the axis over one or multiple consecutive cardiac cycles for the input layer and an associated instance value indicative of the EF for the output layer.