Related Experiment Video
Updated: Dec 11, 2025
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
1.9K
Texture recognition of pulmonary nodules based on volume local direction ternary pattern
Zhipeng Fan1,2, Huadong Sun1,2, Cong Ren1,2
1School of Computer and Information Engineering, Harbin University of Commerce , Harbin, China.
Bioengineered
|August 21, 2020
Summary
This study introduces an AI-assisted diagnostic method for early lung cancer detection using computed tomography (CT) imaging. The novel approach accurately identifies pulmonary nodules, improving diagnostic accuracy and reducing misdiagnosis rates.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Increasing lung cancer incidence necessitates improved early detection methods.
- Manual interpretation of computed tomography (CT) scans for pulmonary nodules is prone to errors like visual fatigue and misdiagnosis.
- Accurate and early identification of pulmonary nodules is critical for effective lung cancer treatment.
Purpose of the Study:
- To develop and validate a novel computer-assisted diagnosis (CAD) system for the accurate detection and classification of pulmonary nodules.
- To enhance the early diagnosis of lung cancer by minimizing human error in CT scan interpretation.
- To provide a reliable tool for assisting radiologists in identifying malignant and benign pulmonary nodules.
Main Methods:
- Image preprocessing and denoising using a median filter.
- Lung parenchyma segmentation via the random walk algorithm to extract regions of interest.
- Texture feature extraction using a volume local direction ternary pattern (VLDT) method on CT slices.
- Pulmonary nodule identification and classification employing a Stacking ensemble algorithm.
Main Results:
- The proposed method achieved an accuracy of 82.2%, sensitivity of 85.7%, and specificity of 78.8% on the LIDC database.
- Cross-validation using ten folds demonstrated the robustness and validity of the diagnostic model.
- The VLDT texture recognition method proved effective for pulmonary nodule identification.
Conclusions:
- The developed Stacking-based CAD system offers a feasible and valuable approach for the early and accurate detection of pulmonary nodules.
- The VLDT feature extraction method shows promise for improving the accuracy of pulmonary nodule recognition in CT imaging.
- This AI-assisted diagnostic tool can serve as a valuable reference for clinicians, potentially reducing misdiagnosis and improving patient outcomes in lung cancer screening.
