Coronary Artery Tissue Classification via OCT and Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing OCT imaging techniques require significant time for physicians to examine coronary artery tissue images, limiting efficiency in tissue type determination.
Innovation Solution
A system utilizing computer-implemented engines, including feature extraction and classification engines with convolutional neural network architectures, to automatically determine coronary artery tissue types from OCT images, employing majority voting for accurate output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple different feature extraction engines with different architectures are used, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system segments the tissue classification task into multiple independent feature extraction engines, each specializing in different architectural approaches (CNN, FCN, hybrid). Each engine processes the OCT image independently to extract specific feature vectors, allowing the system to leverage diverse computational strategies without requiring a single monolithic complex system.
Solution Approach 2:
The system implements a universal classification framework that accepts feature vectors from multiple different engine types (CNN, FCN, hybrid architectures). The classification engine is designed to handle diverse feature representations uniformly, enabling the system to benefit from multiple architectural approaches while maintaining a consistent processing pipeline.
2Measurement precision
If multiple classification engines are used to determine preliminary tissue types, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary classification actions by having multiple classification engines independently determine preliminary tissue types for the region of interest. Each engine processes the feature vectors to generate preliminary classifications, which are then synthesized through majority voting to produce the final tissue type determination.
Solution Approach 2:
The system implements a feedback mechanism through majority voting, where the preliminary classifications from multiple engines are aggregated and compared. The majority vote provides feedback that resolves discrepancies between different engine outputs, ensuring high accuracy while maintaining efficient processing through parallel operation of the classification engines.
Data Source
AI summary
There is described a system for determining a coronal artery tissue type. The system generally has an optical coherence tomography (OCT) imaging system being configured for acquiring an OCT image of coronal artery tissue; and a controller configured to perform the steps of: using trained feature extraction engines, extracting corresponding feature vectors comprising a plurality of features in at least a region of interest of the OCT image; using trained classification engines, determining corresponding preliminary coronal artery tissue types associated to the region of interest of the OCT image based on corresponding ones of the plurality of feature vectors; and using a majority voting engine, majority voting an output coronal artery tissue type associated to the region of interest of the OCT image based on the previously determined preliminary coronal artery tissue types.