CXGNet: A tri-phase chest X-ray image classification for COVID-19 diagnosis using deep CNN with enhanced grey-wolf...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Disease Research

Background:

• The COVID-19 pandemic significantly impacted global health and daily life.
• Early detection of COVID-19, especially in asymptomatic patients, is crucial for reducing mortality and preventing spread.
• Chest X-ray (CXR) imaging is a cost-effective and rapid diagnostic tool for lung conditions, favored over CT scans for certain applications.

Purpose of the Study:

• To develop and evaluate a novel deep learning model for classifying COVID-19 from CXR images.
• To enhance the accuracy and efficiency of COVID-19 diagnosis using artificial intelligence.
• To propose an optimal feature selection technique for improved classification performance.

Main Methods:

• A tri-stage classification model, CXGNet, was developed using deep learning convolutional neural networks (DLCNN).
• An optimal feature selection technique, enhanced grey-wolf optimizer with genetic algorithm (EGWO-GA), was integrated into the model.
• The model was implemented and evaluated for multi-class classification (4-class, 3-class, and 2-class) based on disease categories.

Main Results:

• The proposed CXGNet model demonstrated superior performance compared to conventional methods.
• Achieved classification accuracies of 94.00% for the 4-class model, 97.05% for the 3-class model, and 100% for the 2-class model.
• The integration of EGWO-GA optimized feature selection, leading to enhanced diagnostic accuracy.

Conclusions:

• CXGNet offers a highly accurate and efficient AI-driven solution for COVID-19 classification using CXR images.
• The model shows significant promise for early and precise diagnosis, particularly in resource-limited or emergency settings.
• The study highlights the effectiveness of combining DLCNN with advanced optimization algorithms for medical image analysis in combating infectious diseases.