Automated detection of COVID-19 from CT scan using convolutional neural network

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer-Aided Diagnosis

Background:

• Early and accurate detection of COVID-19 is crucial for pandemic control.
• Artificial intelligence (AI) and deep learning offer potential for automated medical image analysis.

Purpose of the Study:

• To develop and evaluate an AI-based algorithm for detecting COVID-19 from CT scans.
• To compare the performance of VGG16 and ResNet50 architectures using Transfer Learning for COVID-19 detection.

Main Methods:

• Utilized Transfer Learning with VGG16 and ResNet50 architectures.
• Applied data augmentation and fine-tuning techniques to optimize models.
• Conducted stratified 5-fold cross-validation for model robustness assessment.

Main Results:

• Achieved over 99% average accuracy in binary classification (COVID-19 vs. Normal).
• Attained 86.74% (VGG16) and 88.52% (ResNet50) average accuracy in multiclass classification (COVID-19 vs. Normal vs. Pneumonia).

Conclusions:

• The proposed AI models demonstrate superior performance in detecting COVID-19 from CT scans.
• The developed algorithm shows significant potential for automated COVID-19 diagnosis.