EDNC: Ensemble Deep Neural Network for COVID-19 Recognition.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• Accurate and rapid COVID-19 detection is crucial for managing the pandemic.
• Deep learning shows promise in aiding COVID-19 diagnosis but current models lack sufficient accuracy.
• Chest computed tomography (CT) images are vital for identifying COVID-19 pneumonia.

Purpose of the Study:

• To develop and evaluate novel deep learning architectures for accurate and fast COVID-19 detection using chest CT images.
• To improve upon the accuracy of existing COVID-19 recognition models.
• To create an automated system for COVID-19 screening to reduce healthcare professional workload.

Main Methods:

• Proposed three deep learning architectures: F-EDNC, FC-EDNC, and O-EDNC.
• Modified and trained sixteen deep learning neural networks using transfer learning on 2458 chest CT images.
• Developed the Enhanced Deep Neural Convolutional (EDNC) architecture using three modified pre-trained models.

Main Results:

• The F-EDNC model achieved the highest accuracy at 97.75% for COVID-19 recognition.
• FC-EDNC and O-EDNC models demonstrated high accuracy at 97.55% and 96.12%, respectively.
• The developed system significantly outperformed existing COVID-19 recognition models.

Conclusions:

• The proposed EDNC architectures, particularly F-EDNC, offer a significant advancement in accurate and rapid COVID-19 detection from CT scans.
• The automated web application facilitates easy upload of CT scans and provides automatic COVID-19 results.
• This system effectively alleviates the burden on medical professionals for COVID-19 screening.