Related Experiment Video
Updated: Oct 4, 2025
08:05
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
14.4K
COVID-19 Detection Based on Lung Ct Scan Using Deep Learning Techniques
S V Kogilavani1, J Prabhu2, R Sandhiya1
1. Department of Computer Science and Engineering, Kongu Engineering College, Perundurai, Erode 638060, Tamil Nadu, India.
Computational and Mathematical Methods in Medicine
|February 4, 2022
Summary
This study introduces deep learning models for COVID-19 detection using CT scans, addressing RT-PCR kit shortages. The VGG16 model achieved the highest accuracy at 97.68% for distinguishing COVID-19 patients.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Infectious Diseases
Background:
- The COVID-19 pandemic, caused by SARS-CoV-2, presents significant global health challenges.
- Diagnostic limitations, including shortages of real-time reverse transcriptase-polymerase chain reaction (RT-PCR) kits, hinder timely patient management.
- Radiological imaging, particularly computed tomography (CT) scans, offers a valuable alternative for disease detection.
Purpose of the Study:
- To evaluate the efficacy of various deep learning algorithms for detecting COVID-19 from CT scans.
- To identify the most accurate deep learning architecture for COVID-19 diagnosis using CT imaging.
Main Methods:
- A dataset of 3873 CT scan images, labeled as "COVID" and "Non-COVID," was utilized.
- Convolutional Neural Network (CNN) architectures including VGG16, DenseNet121, MobileNet, NASNet, Xception, and EfficientNet were employed.
- The dataset was systematically divided into training, testing, and validation sets for model evaluation.
Main Results:
- VGG16 achieved an accuracy of 97.68%.
- DenseNet121 demonstrated 97.53% accuracy.
- MobileNet, NASNet, Xception, and EfficientNet showed accuracies of 96.38%, 89.51%, 92.47%, and 80.19%, respectively.
- VGG16 outperformed other evaluated architectures in COVID-19 detection accuracy.
Conclusions:
- Deep learning models, particularly VGG16, show high potential for accurate COVID-19 detection using CT scans.
- CT-based deep learning analysis can serve as a viable alternative or supplement to RT-PCR testing.
- Further research can optimize these AI models for widespread clinical application in pandemic scenarios.
