An Automated Early Detection and Classification Method for COVID-19 Stages Based on Deep Learning Technique Using Chest CT Images.

Document Type : Research Studies

Authors

1 Assistant Professor., Computers and Systems Department., Faculty of Engineering., El-Mansoura University., Mansoura., Egypt.

2 Assistant Professor., Department of Electronics and Communications ., Misr Higher Institute for Engineering and Technology., El-Mansoura University., Mansoura., Egypt.

Abstract

Coronavirus Disease 2019 (COVID-19) has widely spread all over the world since the ending of 2019. Until now, the death toll and injuries counted by the world's newest has not stopped. It is better to find an automatic classification technique to find out the extent of pneumonia, and be helpful tools for faster decisions in clinical practice. The Early detection of COVID-19 is an important and urgent need for stopping the spread of the disease. The aim of this study is to develop an automated early detection and classification method for COVID-19 patients based on deep learning technique using chest CT images. The proposed technique classifies the five COVID-19 infection grades with accuracy 98%. It based on using a deep convolutional neural network (CNN) with the ResNet50 model. It detects the early infection grades with a precision 98.3%. Several statistical performance measures had been evaluated for multi-classification categories.  The proposed technique helps the physicians to make faster decisions and treatments for different COVID-19 grades.

Keywords

Main Subjects

Files

Share

How to cite

Statistics