Efficient classification of COVID-19 CT scans by using q-transform model for feature extraction.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Artificial Intelligence in Healthcare

Background:

• Computed Tomography (CT) is crucial for medical imaging and has shown utility in screening for COVID-19 by examining lung appearance.
• Advanced image analysis techniques, driven by technological advancements, enhance medical imaging capabilities.

Purpose of the Study:

• To develop an intelligent and efficient algorithm for classifying COVID-19 cases using CT scans.
• To accurately discriminate between positive and negative COVID-19 cases based on lung image analysis.

Main Methods:

• Image preprocessing using Contrast Limited Adaptive Histogram Equalization (CLAHE) to enhance image details.
• Feature extraction via the q-transform method, measuring pixel grey-level intensity.
• Feature reduction using mean, skewness, and standard deviation, followed by classification using k-nearest neighbor, decision tree, and support vector machine algorithms.

Main Results:

• The proposed algorithm achieved high classification accuracies: 98% for k-nearest neighbor, 98% for decision tree, and 98.25% for support vector machine.
• The experimental results demonstrate the effectiveness of the feature extraction and reduction steps.

Conclusions:

• The developed CT scan-based COVID-19 classification algorithm is efficient, accurate, and flexible.
• The algorithm shows suitability for real-world implementation due to its high classification accuracy across various scenarios.