Cardiac Region Segmentation in CT Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing cardiac regions in CT images, particularly for identifying chronic thromboembolic pulmonary hypertension (CTEPH), lack an efficient automated process for segmenting the cardiac region, including separating the pulmonary artery and aorta from the ventricles.
Innovation Solution
A method involving the receipt of a CT image with contrast enhancement, applying a grey-scale threshold to generate a binary image, followed by objectness filters to create coarse and fine-structure images, fitting these into an atlas of the heart to separate blood vessels and ventricles, and displaying or storing the segmented images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual segmentation methods are used to separate cardiac structures, then segmentation accuracy can be maintained, but diagnostic time and workload increase significantly
Solution Approach 1:
The system performs automated segmentation using contrast enhancement properties and objectness filters, allowing the image processing system to serve itself without requiring manual intervention. The algorithm automatically identifies and separates cardiac structures based on their inherent contrast characteristics.
Solution Approach 2:
The method transforms the segmentation problem by changing parameters - using contrast enhancement properties and applying objectness filters with specific scale parameters to automatically differentiate between blood vessels and ventricles based on their structural characteristics.
2Productivity
If automated segmentation is implemented, then diagnostic efficiency improves, but segmentation precision may deteriorate due to complexity of automatic algorithms
Solution Approach 1:
The objectness filter acts as an intermediary that bridges automated processing and precise segmentation. It processes the contrast-enhanced image to generate a response map that highlights structural boundaries, enabling automated yet precise separation of cardiac structures without manual intervention.
Solution Approach 2:
The method divides the cardiac region into distinct segments (blood vessels and ventricles) based on contrast enhancement patterns and objectness filter responses, allowing automated processing while maintaining precision through structured division of the imaging space.
3Measurement precision
If contrast enhancement is applied to improve structure differentiability, then image quality improves, but image processing complexity and computational requirements increase
Solution Approach 1:
Contrast enhancement is applied preliminarily to the CT image before segmentation processing. This pre-treatment step enhances the differentiability of cardiac structures, making subsequent automated segmentation easier and more accurate without requiring complex processing algorithms.
Data Source
AI summary
The present disclosure pertains to the analysis of the cardiac region in CT images. Provided herein are a method, a computer system and a computer program product for the segmentation of the cardiac region in CT images.