COVID‑19 detection from chest X-ray images using transfer learning.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Diseases

Background:

• COVID-19, a highly contagious coronavirus, poses significant global health risks.
• Current diagnostic methods for COVID-19, like throat and nose swabs, have limitations in sensitivity and are prone to errors.
• Early diagnosis is critical for timely isolation of suspected cases and preventing further transmission.

Purpose of the Study:

• To propose a deep learning framework for early COVID-19 diagnosis using chest X-ray images.
• To enhance the classification performance of Convolutional Neural Network (CNN) models through image pre-processing techniques.
• To evaluate the effectiveness of pre-trained CNN models, specifically VGG19 and EfficientNetB0, for COVID-19 detection.

Main Methods:

• A two-phase framework involving image pre-processing and classification using transfer learning.
• Application of various image enhancement techniques to full and segmented X-ray images.
• Utilized pre-trained VGG19 and EfficientNetB0 models for binary and 4-class classification tasks.

Main Results:

• The VGG19 model, using enhanced full X-ray images, achieved a sensitivity of 0.96, specificity of 0.94, precision of 0.9412, F1 score of 0.9505, and accuracy of 0.95 for binary classification.
• The proposed framework demonstrated a promising classification accuracy of 0.935 for a 4-class classification task.
• Image enhancement techniques improved the classification performance of the CNN models.

Conclusions:

• The developed deep learning framework shows significant potential for the early and accurate diagnosis of COVID-19 from chest X-ray images.
• The use of transfer learning with pre-trained CNN models like VGG19 offers an effective approach for medical image classification tasks.
• This AI-driven approach can support healthcare professionals in making faster diagnostic decisions, thereby improving patient outcomes and public health measures.