Related Experiment Video
Updated: Apr 16, 2026
10:26
Author Spotlight: A 3D Digital Model for the Diagnosis and Treatment of Pulmonary Nodules
Published on: May 19, 2023
2.6K
Pulmonary nodule detection in CT images based on shape constraint CV model
Bing Wang1, Xuedong Tian1, Qian Wang2
1College of Mathematics and Computer Science, Hebei University, Baoding 071002, China.
Medical Physics
|March 5, 2015
Summary
This study introduces a novel method using shape analysis and blood vessel detection to improve pulmonary nodule identification in CT scans, significantly reducing false positives for challenging nodule types.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Radiology
Background:
- Accurate pulmonary nodule detection is crucial for early lung cancer diagnosis.
- Low contrast and adherence to surrounding structures pose significant challenges for computer-aided diagnosis (CADx) systems.
- Existing CADx systems struggle with false positives (FP) due to nodule-like features in adjacent tissues.
Purpose of the Study:
- To develop an enhanced method for pulmonary nodule detection in computed tomography (CT) images.
- To reduce false positive (FP) rates in nodule detection by analyzing shape features and blood vessel adherence.
- To improve the accuracy of CADx systems for challenging nodule types.
Main Methods:
- A three-stage scheme involving lung parenchyma segmentation, candidate nodule extraction, and FP reduction.
- Utilized gray level enhancement and spherical shape filters for candidate nodule extraction.
- Employed a shape-constrained Chan-Vese (CV) model and analysis of adhered blood branches to refine nodule candidates and reduce FPs.
Main Results:
- The proposed method achieved 88% detection of all nodules across 103 cases (127 nodules).
- Effectively identified challenging nodule types: juxta-pleural, juxta-vascular, and ground glass opacity nodules.
- Demonstrated a low false positive rate of 4 FPs per case.
Conclusions:
- The developed method is feasible and effective for detecting challenging pulmonary nodules.
- The integration of shape analysis and vascularity assessment enhances nodule detection accuracy.
- This approach shows promise for improving the reliability of CADx systems in lung nodule diagnosis.
