Related Experiment Video
Updated: Apr 19, 2026
07:53
Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer
Published on: October 13, 2023
2.3K
An automated lung segmentation approach using bidirectional chain codes to improve nodule detection accuracy.
Shiwen Shen1, Alex A T Bui2, Jason Cong3
1Department of Bioengineering, University of California, Los Angeles, CA, USA; Medical Imaging Informatics (MII) Group, Department of Radiological Sciences, University of California, Los Angeles, CA, USA.
Computers in Biology and Medicine
|January 6, 2015
Summary
A new parameter-free lung segmentation algorithm improves nodule detection accuracy, especially for juxtapleural nodules. This method enhances computer-aided detection (CAD) by accurately segmenting lung borders in CT scans.
Area of Science:
- Medical Imaging
- Radiology
- Computer Vision
Background:
- Computer-aided detection and diagnosis (CAD) is crucial for improving radiologist accuracy in lung disease detection.
- Lung segmentation is a necessary preprocessing step for most CAD systems.
- Accurate segmentation is vital for identifying lung nodules, particularly those near the pleura.
Purpose of the Study:
- To propose a parameter-free lung segmentation algorithm.
- To enhance the accuracy of lung nodule detection, with a specific focus on juxtapleural nodules.
- To minimize over-segmentation of adjacent regions while smoothing lung borders.
Main Methods:
- A bidirectional chain coding method was employed.
- A support vector machine (SVM) classifier was integrated for border smoothing.
- The algorithm was tested on 233 computed tomography (CT) studies from the Lung Imaging Database Consortium (LIDC).
Main Results:
- The automated method achieved a 92.6% re-inclusion rate for juxtapleural nodules.
- Segmentation accuracy was validated on 10 CT series.
- An average over-segmentation ratio of 0.3% and an under-segmentation rate of 2.4% were observed compared to manual segmentation.
Conclusions:
- The proposed parameter-free lung segmentation algorithm effectively improves lung nodule detection accuracy.
- The method demonstrates high accuracy in segmenting lung borders, particularly for juxtapleural nodules.
- This technique offers a valuable tool for enhancing CAD systems in thoracic imaging analysis.
