COVID-19 pneumonia level detection using deep learning algorithm and transfer learning

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Radiology

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
• Assessing lung inflammation severity is crucial for patient management.
• Current diagnostic methods can be time-consuming.

Purpose of the Study:

• To develop an AI engine for classifying COVID-19 lung inflammation levels (mild, progressive, severe).
• To evaluate the performance of AI models in analyzing CT scan images for pneumonia detection.

Main Methods:

• A two-phase approach using morphological analysis for lesion quantification.
• Classification using a modified Convolution Neural Network (CNN) and k-Nearest Neighbors (KNN).
• Transfer learning was employed to train the CNN model on smaller datasets.

Main Results:

• The modified CNN model achieved up to 95.65% testing accuracy in classifying lung inflammation.
• The CNN model demonstrated superior performance compared to other classification algorithms.
• Transfer learning enabled the CNN to achieve 92.80% accuracy for pneumonia detection on X-ray images.

Conclusions:

• The developed AI engine effectively classifies COVID-19 lung inflammation severity from CT scans.
• The system shows potential for efficient application to both CT and X-ray image datasets.
• AI-powered image analysis offers a promising avenue for rapid pandemic diagnosis.