Hybrid PSO-SVM algorithm for Covid-19 screening and quantification.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Diseases

Background:

• The COVID-19 pandemic significantly increased the diagnostic burden on radiologists.
• Accurate and timely diagnosis is crucial for managing outbreaks and patient care.

Purpose of the Study:

• To develop and evaluate an AI algorithm for automated analysis of Computed Tomography (CT) images to detect COVID-19-related pneumonia.
• To reduce physician workload and improve diagnostic efficiency in outbreak scenarios.

Main Methods:

• A hybrid Particle Swarm Optimization-Support Vector Machine (PSO-SVM) AI algorithm was developed.
• The algorithm was trained to analyze CT images and classify the presence of pneumonia.
• The system was deployed globally, and challenges like data security and model effectiveness were addressed.

Main Results:

• The AI model achieved a specificity of 0.85, sensitivity of 0.956, and accuracy of 95.78%.
• The system demonstrated the ability to positively handle challenges such as data security and testing time.
• Early identification of infected patients facilitated timely confirmation and segregation.

Conclusions:

• The proposed AI-based system effectively assists in the diagnosis of COVID-19 pneumonia from CT scans.
• This technology has the potential to save approximately 50% of physicians' time, particularly in outbreak settings.
• The AI integrated system offers a reliable solution for improving diagnostic speed and accuracy during global health crises.