Covid-19: automatic detection from X-ray images utilizing transfer learning with convolutional neural networks.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Disease Diagnostics

Background:

• Accurate and timely diagnosis of Coronavirus disease (COVID-19) is critical for patient management and public health.
• Traditional diagnostic methods for COVID-19 have limitations, including varying failure rates.
• X-ray imaging presents a potential alternative or supplementary diagnostic tool.

Purpose of the Study:

• To evaluate the performance of state-of-the-art convolutional neural network (CNN) architectures for automatic COVID-19 detection using X-ray images.
• To assess the efficacy of Transfer Learning techniques in medical image classification for identifying COVID-19 and other pneumonias.
• To determine the diagnostic potential of deep learning models applied to chest X-rays for COVID-19.

Main Methods:

• Utilized two datasets comprising X-ray images of patients with COVID-19, common bacterial pneumonia, and normal conditions.
• Employed Transfer Learning, a deep learning procedure, to train CNN architectures on medical image classification tasks.
• Collected data from publicly available medical image repositories.

Main Results:

• Deep learning models demonstrated the ability to extract significant biomarkers for COVID-19 from X-ray images.
• The best performance achieved was 96.78% accuracy, 98.66% sensitivity, and 96.46% specificity.
• The models effectively differentiated between COVID-19, bacterial pneumonia, and normal conditions.

Conclusions:

• Deep learning applied to X-ray imaging is a viable method for detecting COVID-19 with high performance.
• The findings suggest that X-rays could be integrated into the diagnostic workflow for COVID-19.
• Further research is recommended to explore the multifaceted aspects of X-ray-based COVID-19 diagnosis.