EffiCOVID-net: A highly efficient convolutional neural network for COVID-19 diagnosis using chest X-ray imaging

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer Science

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools, particularly for respiratory illnesses.
• Chest X-ray analysis is vital for differentiating COVID-19 from other conditions.
• Deep learning models show promise in medical image diagnostics but often face challenges with computational intensity and limited data.

Purpose of the Study:

• To develop and evaluate an efficient deep learning model, EffiCOVID-Net, for the accurate detection of COVID-19 from chest X-ray images.
• To address limitations of existing models, such as computational overhead and inability to capture multi-scale features.

Main Methods:

• Proposed EffiCOVID-Net, a convolutional neural network (CNN) utilizing diverse feature learning units and recurrent connections within EffiCOVID blocks.
• Employed (3x3) convolution filters to extract complex features while maintaining spatial integrity.
• Evaluated model performance on two public COVID-19 chest X-ray datasets using standard metrics and Grad-CAM for visualization.

Main Results:

• EffiCOVID-Net achieved high accuracy in multi-class classification (COVID-19 vs. Normal vs. Pneumonia): 98.68% (D1), 98.55% (D2), 98.87% (DMix).
• Attained superior accuracy in binary classification (COVID-19 vs. Normal): 99.06% (D1), 99.78% (D2), 99.07% (DMix).
• Demonstrated superior accuracy, efficiency, and robustness compared to existing methods, with enhanced interpretability via Grad-CAM.

Conclusions:

• EffiCOVID-Net is a highly efficient and accurate deep learning model for COVID-19 detection from chest X-rays.
• Its lightweight architecture makes it suitable for resource-constrained clinical environments.
• The model serves as a valuable assistive tool, enhancing diagnostic workflows but should not replace clinical judgment.