A novel deep neuroevolution-based image classification method to diagnose coronavirus disease (COVID-19)

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• COVID-19 significantly disrupts public health and the global economy.
• Rapid diagnosis of COVID-19 is crucial for early patient management and community containment.
• Chest X-ray imaging offers vital diagnostic information for COVID-19 detection.

Purpose of the Study:

• To propose a novel deep neuroevolution (DNE) framework for automated COVID-19 diagnosis from chest X-ray images.
• To enhance the accuracy and efficiency of deep learning (DL) model design for medical image analysis.
• To address the shortage of medical expertise in rural areas through automated diagnostic tools.

Main Methods:

• Development of a novel two-stage improved deep neuroevolution (DNE) algorithm.
• Application of DNE techniques for automated design of deep learning (DL) architectures.
• Evaluation of the proposed DNE framework on a real-world chest X-ray dataset.

Main Results:

• The proposed DNE framework demonstrated superior classification performance.
• The system achieved high accuracy across various evaluation metrics for COVID-19 detection.
• Automated DL architecture design via DNE improved diagnostic capabilities.

Conclusions:

• The novel DNE algorithm offers an effective approach for automated COVID-19 diagnosis using chest X-rays.
• This method can improve diagnostic accuracy and support medical professionals, especially in underserved regions.
• Deep neuroevolution presents a promising avenue for optimizing AI in medical diagnostics.