Related Experiment Video
Updated: May 23, 2026
Three-Dimensional Phase Resolved Functional Lung Magnetic Resonance Imaging
Published on: June 21, 2024
Automatic left ventricle segmentation in cardiac MRI using topological stable-state thresholding and region
Hong Liu1, Huaifei Hu, Xiangyang Xu
1Center for Biomedical Imaging and Bioinformatics, Key Laboratory of Education Ministry for Image Processing and Intelligence Control, School of Computer Science and Technology, Huazhong University of Science and Technology, 1037 Luo Yu Road, Wuhan, Hubei, China.
This study introduces an advanced algorithm for segmenting the left ventricle (LV) in cardiac MRI scans, improving accuracy for cardiovascular disease diagnosis. The novel method enhances automated segmentation of the LV, crucial for assessing cardiac function.
Area of Science:
- Medical Imaging
- Cardiovascular Science
- Image Processing
Background:
- Accurate left ventricle (LV) segmentation is vital for assessing cardiac function parameters using cardiac magnetic resonance imaging (MRI).
- Existing automated segmentation methods may lack the robustness and accuracy required for clinical applications.
Purpose of the Study:
- To develop a novel and robust algorithm for automatic LV segmentation on short-axis cardiac MRI.
- To enhance the accuracy of LV segmentation for improved cardiac functional parameter assessment.
Main Methods:
- A dataset of 45 cardiac MRI cases (including normal, heart failure, and hypertrophy) was utilized.
- The algorithm incorporates topological stable-state thresholding for endocardial contour refinement.
- It also employs an edge map with non-maxima gradient suppression and region-restricted dynamic programming for epicardial boundary derivation.
Main Results:
- The algorithm achieved approximately 91% good contours for both endocardial and epicardial boundaries.
- Average perpendicular distance was ~2 mm, with an overlapping Dice metric of ~0.91.
- High correlation (R²=0.9048 for ejection fraction, R²=0.8221 for LV mass) was observed between the automated method and expert assessments.
Conclusions:
- A novel automatic method for LV segmentation in cardiac MRI has been developed using topological stable-state thresholding and region-restricted dynamic programming.
- The proposed method demonstrates improved accuracy and robustness in LV segmentation.
- This approach holds significant potential for enhancing computer-aided diagnosis systems in cardiovascular diseases.
