COVID-19 image classification using deep features and fractional-order marine predators algorithm.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
• Convolutional Neural Networks (CNNs) excel at image classification but require significant computational resources.
• Feature selection is crucial for optimizing CNN performance and reducing complexity.

Purpose of the Study:

• To develop an efficient hybrid classification model for COVID-19 detection from X-ray images.
• To reduce the computational cost associated with deep learning models for medical image analysis.
• To improve the accuracy and performance of COVID-19 image classification.

Main Methods:

• A hybrid approach combining Inception CNN for feature extraction and the Marine Predators Algorithm (MPA) for feature selection.
• Integration of fractional-order calculus (FO) with MPA, termed FO-MPA, for enhanced feature selection.
• Evaluation on two publicly available COVID-19 X-ray datasets.

Main Results:

• The FO-MPA approach significantly reduced the number of features, selecting 130 and 86 features from over 51K.
• Achieved high classification accuracy, reaching 98.7% and 98.2% for dataset 1, and 99.6% and 99% for dataset 2 (accuracy and F-Score).
• Outperformed several existing CNN models and recent works on COVID-19 image classification.

Conclusions:

• The proposed FO-MPA hybrid model offers a computationally efficient and highly accurate method for COVID-19 detection.
• This approach effectively balances high performance with reduced computational complexity in medical image analysis.
• The study demonstrates the potential of integrating swarm intelligence and fractional calculus for improved diagnostic AI.