Related Experiment Video
Updated: May 9, 2026
08:05
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
Automated iterative neutrosophic lung segmentation for image analysis in thoracic computed tomography
Yanhui Guo1, Chuan Zhou, Heang-Ping Chan
1Department of Radiology, University of Michigan, Ann Arbor, Michigan 48109, USA. yguo@stu.edu
Medical Physics
|August 10, 2013
Summary
Accurate lung segmentation in CT scans is crucial for disease detection. The new iterative neutrosophic lung segmentation (INLS) method improves upon previous techniques, especially for lungs affected by disease, enhancing computer-aided detection systems.
Area of Science:
- Medical Imaging
- Image Analysis
- Radiology
Background:
- Lung segmentation in thoracic computed tomography (CT) is vital for analyzing lung diseases and abnormalities.
- Previous expectation-maximization and morphological operations (EMM) method faced challenges with extensive lung diseases.
- Accurate segmentation is essential for computer-aided detection (CAD) systems, such as for pulmonary embolism (PE).
Purpose of the Study:
- To develop an improved lung segmentation method capable of accurately segmenting lungs, even in the presence of significant disease.
- To refine the existing expectation-maximization and morphological operations (EMM) method for enhanced accuracy in thoracic CT imaging.
Main Methods:
- Developed an iterative neutrosophic lung segmentation (INLS) method, refining initial lung regions (ILRs) obtained from EMM.
- Utilized anatomic features of ribs and lungs, mapping them into a neutrosophic domain for iterative refinement of ILRs.
- Evaluated performance on 58 CTPA scans (34 with lung disease) using Percentage Overlap Area (POA), Hausdorff distance (Hdist), and Average Distance (AvgDist) against radiologist-outlined regions.
Main Results:
- The INLS method significantly improved segmentation accuracy compared to the EMM method.
- Average POA increased from 85.4±18.4% (EMM) to 91.2±6.7% (INLS); Hdist decreased from 22.6±29.4 mm to 16.0±11.3 mm; AvgDist decreased from 3.5±5.4 mm to 2.5±1.0 mm (p<0.05).
- INLS demonstrated superior performance over EMM in segmenting lungs with and without disease, particularly excelling in diseased cases.
Conclusions:
- The iterative neutrosophic lung segmentation (INLS) method significantly enhances lung segmentation accuracy, especially in diseased thoracic CT images.
- This improved segmentation facilitates downstream image analysis tasks and computer-aided detection (CAD) applications for lung conditions.
- The method shows promise for improving automated pulmonary embolism detection in CT pulmonary angiography (CTPA).
