COVID-19 CT-images diagnosis and severity assessment using machine learning algorithm.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• Current COVID-19 diagnostic tools like RT-PCR have limitations in assessing disease severity.
• Accurate staging of COVID-19 is crucial for effective treatment and patient management.

Purpose of the Study:

• To develop a simple, reliable, and automatic system for diagnosing COVID-19 severity from CT scans.
• To classify COVID-19 severity into mild, moderate, and severe stages.

Main Methods:

• Utilized a dataset of 1801 CT scans for training and validation.
• Employed image segmentation and feature extraction (infection ratio, statistical textures, GLCM, GLRLM).
• Applied four machine learning algorithms: Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Decision Trees (DT), and Naïve Bayes for classification.

Main Results:

• The SVM model achieved high accuracy: 99.12% (normal), 98.24% (mild), 98.73% (moderate), and 99.9% (severe).
• The model demonstrated a high Area Under the Curve (AUC) of 0.99.
• The proposed system outperformed existing state-of-the-art models in COVID-19 severity classification.

Conclusions:

• The developed automated system effectively diagnoses COVID-19 severity from CT scans.
• This tool can assist clinicians in determining appropriate treatment dosages and timelines.
• The system offers a significant improvement over current diagnostic methods for assessing COVID-19 progression.