Refining COVID-19 Lesion Segmentation in Lung CT Scans Using Swarm Intelligence and Evolutionary Algorithms.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Accurate identification of lung lesions in Computed Tomography (CT) scans is critical for managing Coronavirus Disease 2019 (COVID-19).
• Swarm intelligence algorithms show potential for automated lesion segmentation in medical imaging.

Purpose of the Study:

• To evaluate and compare the performance of four swarm intelligence algorithms for segmenting COVID-19 related lung lesions in CT scans.
• To determine the most effective algorithm for improving diagnostic accuracy in COVID-19 detection.

Main Methods:

• The study implemented and compared four swarm intelligence algorithms: Gravitational Search Algorithm (GSA), Bacterial Foraging Optimization Algorithm (BFOA), Genetic Algorithm (GA), and Particle Swarm Optimization (PSO).
• These algorithms were applied to segment lung lesions indicative of COVID-19 in CT images.

Main Results:

• Genetic Algorithm (GA), Gravitational Search Algorithm (GSA), and Bacterial Foraging Optimization Algorithm (BFOA) demonstrated segmentation accuracies above 90.5%.
• Particle Swarm Optimization (PSO) achieved the highest segmentation accuracy at 91.45% with an F1 score of 95.54%.
• The overall segmentation accuracy reached up to 99% using the optimized swarm intelligence approach.

Conclusions:

• Swarm and evolutionary algorithms are effective tools for segmenting COVID-19 lung lesions.
• The study highlights the potential of these algorithms to enhance diagnostic accuracy and treatment efficiency for COVID-19 patients.