CORONA-Net: Diagnosing COVID-19 from X-ray Images Using Re-Initialization and Classification Networks

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
• Accurate testing for COVID-19 remains a significant challenge.
• Convolutional Neural Networks (CNNs) offer potential for automated analysis of medical images.

Purpose of the Study:

• To develop and evaluate a novel CNN model, CORONA-Net, for detecting COVID-19 from chest X-ray images.
• To improve the accuracy and efficiency of COVID-19 diagnosis through an automated system.
• To provide radiologists with a tool for validating their diagnostic findings.

Main Methods:

• A two-phase CNN architecture, CORONA-Net, was designed, comprising a reinitialization phase (encoder-decoder) and a classification phase (encoder backbone).
• The reinitialization phase trained encoder and decoder networks using medical image distributions.
• The classification phase fine-tuned the encoder network using weights from the reinitialization phase.

Main Results:

• CORONA-Net achieved a high overall accuracy of 95.84% in detecting COVID-19 from chest X-rays.
• The proposed network demonstrated superior performance compared to existing state-of-the-art methods.
• Extensive experiments were conducted on the publicly available COVIDx dataset.

Conclusions:

• CORONA-Net is a highly accurate and effective deep learning model for COVID-19 detection using chest X-rays.
• The two-phase approach of CORONA-Net enhances diagnostic capabilities.
• This automated system shows promise in combating the COVID-19 pandemic through improved early detection.