Automated System for Identifying COVID-19 Infections in Computed Tomography Images Using Deep Learning Models

Area of Science:

• Medical Imaging and Artificial Intelligence

Background:

• Coronavirus disease 2019 (COVID-19) presents a significant global health challenge with a high fatality rate.
• Current diagnostic methods combining clinical, epidemiological, and laboratory data can be insufficient, potentially leading to missed cases and complications.
• Deep learning offers a promising approach for enhancing medical image analysis in diagnostics.

Purpose of the Study:

• To develop and evaluate a deep learning-based diagnostic system for COVID-19 detection using computed tomography (CT) images.
• To compare the performance of Convolutional Neural Network (CNN), stacked autoencoder, and deep neural network models in identifying COVID-19 cases from CT scans.

Main Methods:

• Utilized a large-scale dataset of CT images for training deep learning models.
• Implemented and adapted CNN, stacked autoencoder, and deep neural network architectures for COVID-19 classification.
• Modified classification procedures prior to applying deep learning techniques to CT images.

Main Results:

• The Convolutional Neural Network (CNN) model demonstrated the highest performance, achieving an accuracy of 88.30%.
• The CNN model also reported a sensitivity of 87.65% and a specificity of 87.97%.
• The proposed deep learning system outperformed existing state-of-the-art models in detecting COVID-19 from CT images.

Conclusions:

• Deep learning models, particularly CNNs, show significant potential for accurate and reliable COVID-19 diagnosis using CT imaging.
• The developed system offers an improved approach to COVID-19 detection, potentially aiding early diagnosis and patient management.
• The study highlights the effectiveness of AI in medical image analysis for infectious disease identification.