Endoscopic Blood Flow Classification Through Reference-Region Color Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for blood flow confirmation in medical imaging, such as the ICG fluorescence method and oxygen saturation imaging, are invasive and lack robustness against external factors like inter-patient differences and light sources, making it difficult to distinguish between normal and ischemic regions accurately.
Innovation Solution
A medical image recognition apparatus using a trained classifier to identify normal and ischemic regions based on color information, employing a convolutional neural network (CNN) for two-class classification, which sets reference and evaluation regions and determines blood flow attributes robustly against external factors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the ICG fluorescence method is used for blood flow confirmation, then the presence or absence of blood flow can be visualized, but it is invasive due to intravenous injection of the drug and requires waiting for infiltration
Solution Approach 1:
The patent extracts and analyzes color information directly from the medical image itself, removing the need for external contrast agents like ICG. The system identifies blood flow attributes by processing the inherent color data in the image, thereby eliminating invasive injections and waiting periods while maintaining blood flow visualization capability
Solution Approach 2:
The patent replaces the chemical/biological mechanism of ICG fluorescence with an optical/image processing mechanism. Instead of using fluorescent dyes that require injection and infiltration time, the system uses color analysis of the captured medical image to determine blood flow status, substituting a non-invasive optical method for the invasive chemical method
2Measurement precision
If oxygen saturation imaging is used to determine blood flow, then the presence or absence of blood flow can be visualized, but it lacks robustness against external factors such as inter-patient differences and light sources
Solution Approach 1:
The patent transforms the approach by changing from absolute oxygen saturation measurement to relative color comparison. Instead of relying on fixed oxygen saturation thresholds that are sensitive to external factors, the system compares color information between different regions within the same image, making the measurement robust against variations in lighting conditions and inter-patient differences
Solution Approach 2:
The patent implements a feedback mechanism by using a trained classifier that learns from training data to identify patterns in color information. The classifier is trained to recognize blood flow attributes based on color characteristics, providing a robust decision-making system that adapts to variations in imaging conditions and patient characteristics
3Difficulty of detecting and measuring
If chromaticities are used to detect abnormal findings, then abnormal regions can be identified, but normal region and ischemic region are not classified
Solution Approach 1:
The patent applies local quality analysis by comparing color information between specific regions (reference region and evaluation region) rather than analyzing the entire image uniformly. This localized comparison enables the system to classify regions as normal or ischemic based on their relative color characteristics, providing both abnormal detection and classification information
Data Source
AI summary
A processor configured to acquire a medical image; set a reference region and an evaluation region different from the reference region in an organ portion of the medical image; input color information of the medical image of the reference region and the medical image of the evaluation region to a trained classifier; output an evaluation result of a blood flow attribute of the evaluation region with respect to a blood flow attribute of the reference region; and perform control to display an evaluation result of the blood flow attribute of the evaluation region.


