Improved COVID-19 detection with chest x-ray images using deep learning.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Diseases

Background:

• The novel coronavirus disease (COVID-19) emerged as a global public health crisis.
• Rapid and accurate diagnosis is crucial for managing COVID-19 spread and reducing healthcare system stress.
• Traditional diagnostic methods can be time-consuming, necessitating faster alternatives.

Purpose of the Study:

• To develop a computer-aided design (CAD) system for classifying chest X-rays into COVID-19, viral pneumonia, or healthy categories.
• To evaluate the performance of pre-trained deep neural networks (DNNs) for this classification task.
• To address the challenge of limited COVID-19 positive chest X-ray datasets.

Main Methods:

• Utilized a dataset of 2905 chest X-ray images (219 COVID-19, 1341 healthy, 1345 viral pneumonia).
• Employed four pre-trained deep neural networks (DNNs) for classification.
• Evaluated models on a test set of 30 images per class.

Main Results:

• AlexNet demonstrated superior performance, achieving an accuracy of 97.6%.
• AlexNet recorded an average precision of 0.98, recall of 0.97, and F1 score of 0.98.
• The study identified an effective DNN for rapid chest X-ray-based disease classification.

Conclusions:

• Deep neural networks, particularly AlexNet, show significant promise for rapid and accurate COVID-19 diagnosis using chest X-rays.
• The developed CAD system can aid healthcare professionals in timely disease identification.
• This approach offers a potential solution to accelerate diagnosis and alleviate pressure on healthcare systems during pandemics.