Related Experiment Video
Updated: Aug 10, 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
1.5K
Lung Diseases Detection Using Various Deep Learning Algorithms
M Jasmine Pemeena Priyadarsini1, Ketan Kotecha2,3, G K Rajini4
1School of Electronics Engineering, Vellore Institute of Technology, Vellore 632014, India.
Journal of Healthcare Engineering
|February 13, 2023
Summary
This study introduces a deep learning framework for detecting pneumonia, tuberculosis, and lung cancer from medical images. The proposed sequential model achieved high accuracy, offering a faster and more effective method for disease diagnosis.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Healthcare
- Computational Pathology
Background:
- Early and accurate detection of lung diseases like pneumonia, tuberculosis, and lung cancer is crucial for effective patient treatment.
- Traditional diagnostic methods can be time-consuming and may face limitations with complex imaging data.
- Deep learning, particularly Convolutional Neural Networks (CNNs), shows significant promise in analyzing biomedical image datasets.
Purpose of the Study:
- To develop and validate a deep learning framework for detecting and classifying multiple lung diseases from X-ray and CT scan images.
- To implement and compare the performance of Sequential, Functional, and Transfer deep learning models for lung disease classification.
- To establish a novel approach for disease detection that potentially outperforms existing methods.
Main Methods:
- Implementation of three deep learning models: Sequential, Functional, and Transfer learning.
- Training of models on open-source datasets of X-ray and CT scan images.
- Validation and performance comparison of the implemented models against existing methods using metrics like accuracy, F1 score, and recall.
Main Results:
- The sequential model achieved high performance for pneumonia (F1 score 98.55%, accuracy 98.43%) and tuberculosis (F1 score 97.99%, accuracy 99.4%).
- The functional model demonstrated superior performance for lung cancer detection with 99.9% accuracy and 99.89% specificity.
- The proposed models offer high accuracy and efficiency, with the functional model requiring fewer computational resources.
Conclusions:
- The developed deep learning framework provides an effective and accurate method for classifying lung diseases from medical images.
- The sequential and functional models represent a significant advancement in automated lung disease detection, offering improved diagnostic capabilities.
- This research paves the way for faster, more cost-effective, and accurate diagnosis of critical lung conditions, benefiting patient care.
