Machine learning based COVID -19 disease recognition using CT images of SIRM database

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• The COVID-19 pandemic necessitated rapid diagnostic tools.
• Computational Tomography (CT) scans show viral infection signals but can be challenging for radiologists to interpret quickly.
• Automated analysis of medical images can improve diagnostic speed and accuracy.

Purpose of the Study:

• To develop and evaluate a machine learning model for accurate and efficient COVID-19 detection from CT images.
• To compare the performance of various feature extraction and machine learning techniques for COVID-19 classification.
• To identify the optimal combination of feature extraction and classification for COVID-19 diagnosis.

Main Methods:

• A dataset of 3764 COVID-19 CT images was curated.
• Feature extraction was performed using Grey Level Co-occurrence Matrix (GLCM) and Discrete Wavelet Transform (DWT).
• Machine learning algorithms including Support Vector Machines (SVM), Linear Discriminant Analysis (LDA), Naive Bayes, K-Nearest Neighbours, and Random Forests were applied for classification.

Main Results:

• The Support Vector Machine (SVM) model combined with Grey Level Co-occurrence Matrix (GLCM) for feature extraction demonstrated superior performance.
• This combination achieved 99.70% accuracy, 99.80% sensitivity, and 97.03% F-score using 10-fold cross-validation.
• The model's effectiveness was consistent across different datasets, indicating robustness.

Conclusions:

• Machine learning, particularly SVM with GLCM feature extraction, offers a highly accurate and efficient method for COVID-19 detection from CT scans.
• This approach can significantly aid radiologists in the rapid diagnosis and management of COVID-19.
• The findings support the integration of AI-powered tools in medical imaging for pandemic response.