Explainable artificial intelligence-based edge fuzzy images for COVID-19 detection and identification.

Area of Science:

• Artificial Intelligence in Medical Imaging
• Deep Learning for Disease Detection
• Radiology and Diagnostic Imaging

Background:

• The COVID-19 pandemic necessitates effective screening methods beyond traditional laboratory tests.
• Chest radiography is a crucial radiological screening tool for identifying COVID-19 infections.
• Limitations in current diagnostic tests highlight the need for advanced screening technologies.

Purpose of the Study:

• To develop and evaluate a Multi-Input Transfer Learning COVID-Net fuzzy convolutional neural network for COVID-19 detection using chest X-ray images.
• To leverage explainability methods to understand critical factors in COVID-19 diagnosis from X-rays.
• To improve the accuracy and efficiency of COVID-19 screening through advanced AI.

Main Methods:

• Implementation of a Multi-Input Transfer Learning COVID-Net fuzzy convolutional neural network.
• Utilizing transfer learning and pre-trained models for enhanced detection accuracy.
• Employing explainability techniques to analyze network predictions and identify key diagnostic features.

Main Results:

• The developed model achieved high accuracy in detecting COVID-19 from X-ray images.
• Achieved an Area Under the Curve (AUC) of 1.0, with accuracy, precision, and recall of 0.97.
• Quantized model for Internet of Things (IoT) devices maintained 0.95% accuracy.

Conclusions:

• Transfer learning and deep learning approaches, like the proposed COVID-Net, are highly effective for COVID-19 detection via chest X-rays.
• The AI model provides valuable insights for clinicians, potentially improving screening protocols.
• The quantized model demonstrates feasibility for deployment on edge devices for real-time screening.