Related Experiment Video
Updated: Aug 9, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Semi-automated three-dimensional segmentation for cardiac CT images using deep learning and randomly distributed
Ted Shi1, Maysam Shahedi1,2, Kayla Caughlin1
1Department of Bioengineering, The University of Texas at Dallas, Richardson, TX.
Insights
This study introduces a semi-automated deep learning method for segmenting cardiac chambers in CT scans. The point-guided approach improves accuracy and efficiency, offering a promising alternative to manual segmentation for cardiovascular disease analysis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Cardiology
Background:
- Cardiovascular diseases (CVDs) necessitate accurate heart segmentation in cardiac CT scans.
- Manual segmentation is laborious and prone to observer variability, impacting results.
- Fully automated deep learning methods for cardiac segmentation still require improvement to match expert accuracy.
Purpose of the Study:
- To develop a semi-automated deep learning approach for cardiac segmentation that balances accuracy and efficiency.
- To bridge the gap between manual segmentation's precision and fully automated methods' speed.
- To enhance the delineation of heart chambers in CT images.
Main Methods:
- A semi-automated deep learning strategy using a fixed number of points on the cardiac surface was employed.
- Points-distance maps were generated from user-selected points.
- A three-dimensional (3D) fully convolutional neural network (FCNN) was trained on these maps for segmentation prediction.
Main Results:
- The method achieved Dice scores ranging from 0.742 to 0.917 across the four heart chambers.
- Average Dice scores were 0.846 ± 0.059 (left atrium), 0.857 ± 0.052 (left ventricle), 0.826 ± 0.062 (right atrium), and 0.824 ± 0.062 (right ventricle).
- The point-guided, image-independent approach demonstrated strong performance in chamber-by-chamber segmentation.
Conclusions:
- The developed semi-automated deep learning method shows promising performance for cardiac segmentation in CT images.
- This approach offers a viable alternative to manual segmentation, improving efficiency without sacrificing significant accuracy.
- The point-guided technique facilitates accurate delineation of individual heart chambers.
Abstract:
Given the prevalence of cardiovascular diseases (CVDs), the segmentation of the heart on cardiac computed tomography (CT) remains of great importance. Manual segmentation is time-consuming and intra-and inter-observer variabilities yield inconsistent and inaccurate results. Computer-assisted, and in particular, deep learning approaches to segmentation continue to potentially offer an accurate, efficient alternative to manual segmentation. However, fully automated methods for cardiac segmentation have yet to achieve accurate enough results to compete with expert segmentation. Thus, we focus on a semi-automated deep learning approach to cardiac segmentation that bridges the divide between a higher accuracy from manual segmentation and higher efficiency from fully automated methods. In this approach, we selected a fixed number of points along the surface of the cardiac region to mimic user interaction. Points-distance maps were then generated from these points selections, and a three-dimensional (3D) fully convolutional neural network (FCNN) was trained using points-distance maps to provide a segmentation prediction. Testing our method with different numbers of selected points, we achieved a Dice score from 0.742 to 0.917 across the four chambers. Specifically. Dice scores averaged 0.846 ± 0.059, 0.857 ± 0.052, 0.826 ± 0.062, and 0.824 ± 0.062 for the left atrium, left ventricle, right atrium, and right ventricle, respectively across all points selections. This point-guided, image-independent, deep learning segmentation approach illustrated a promising performance for chamber-by-chamber delineation of the heart in CT images.
