Related Experiment Video
Updated: Mar 10, 2026
Author Spotlight: Optimized Lung MRI Protocol with Computationally Efficient Reconstruction Methods
Published on: September 6, 2024
Anatomy-based algorithm for automatic segmentation of human diaphragm in noncontrast computed tomography images
Elham Karami1, Yong Wang2, Stewart Gaede3
1Western University, Department of Medical Biophysics, Medical Sciences Building, London, Ontario N6A 5C1, Canada; Robarts Research Institute, Imaging Research Laboratories, 1151 Richmond Street, London, Ontario N6A 5B7, Canada.
This study introduces an automatic algorithm for segmenting the diaphragm in 4-D CT scans, overcoming challenges like low contrast and motion artifacts. The novel method accurately delineates the entire diaphragm, aiding respiratory system research.
Area of Science:
- Medical Imaging
- Anatomy
- Physiology
Background:
- The diaphragm is crucial for respiration, making its anatomical and physiological study important.
- Four-dimensional (4-D) computed tomography (CT) offers valuable data but presents segmentation challenges.
- Diaphragm segmentation is difficult due to low image contrast and motion artifacts in 4-D CT.
Purpose of the Study:
- To develop an automatic algorithm for segmenting the diaphragm in 4-D CT images.
- To address limitations of current segmentation methods, including lack of contrast and motion artifacts.
- To enable comprehensive analysis of diaphragm anatomy and function.
Main Methods:
- An automatic segmentation algorithm leveraging a priori anatomical knowledge.
- Utilizing surrounding organs (lungs, heart, aorta, ribcage) to guide diaphragm segmentation.
- A novel approach to overcome segmentation difficulties in 4-D CT scans.
Main Results:
- The algorithm achieved favorable results with an average mean distance of [Formula: see text] compared to manual segmentation.
- Successfully delineated the entire diaphragm, a first for segmentation algorithms.
- Demonstrated the potential for accurate diaphragm boundary condition acquisition.
Conclusions:
- The proposed algorithm effectively segments the diaphragm in 4-D CT images.
- This technique overcomes significant challenges in medical image analysis.
- Enables advanced applications like biomechanical modeling for diaphragm physiology research.
