COVID-19 detection in CT and CXR images using deep learning models.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Disease Diagnostics

Background:

• Infectious diseases like COVID-19 pose significant global health risks.
• Rapid and accurate diagnosis is crucial for managing pandemics.
• Medical imaging, including CT scans and Chest X-rays (CXR), offers a viable diagnostic approach.

Purpose of the Study:

• To investigate the efficacy of deep transfer learning models for detecting COVID-19.
• To evaluate the performance of various pre-trained deep neural networks using CT and CXR images.
• To assess the impact of data augmentation on model generalization.

Main Methods:

• Utilized deep transfer learning with pre-trained models: ResNet50, InceptionV3, VGGNet-19, and Xception.
• Employed data augmentation to expand the training dataset and prevent overfitting.
• Evaluated model performance on separate CT and CXR image datasets, as well as combined modalities.

Main Results:

• VGGNet-19 achieved 87% accuracy on CT scans, while Xception reached 98% accuracy on CXR images.
• Combined analysis of CT and X-ray data using VGG-19 yielded 90.5% accuracy.
• Deep learning models demonstrated strong performance in identifying COVID-19 positive cases.

Conclusions:

• Deep transfer learning is a powerful tool for automated COVID-19 detection using medical imaging.
• The study highlights the potential of AI in enhancing diagnostic capabilities, especially in resource-limited settings.
• Specific models like VGGNet-19 and Xception show promise for clinical application in COVID-19 screening.