Intelligent tool for detecting Covid-19 using convolutional neural network based on both CT and x-ray lung images

المجلة: AIP Conference Proceedings 2ND INTERNATIONAL CONFERENCE ON ENGINEERING & SCIENCE

سنة النشر: 2021

تاريخ النشر: 2024-07-10

0

Intelligent tool for detecting Covid-19 using convolutional neural network based on both CT and x-ray lung images

المجلة: 2ND INTERNATIONAL CONFERENCE ON ENGINEERING & SCIENCE

سنة النشر: 2021

تاريخ النشر: 2024-07-11

The pandemic COVID19 that has been emerged around the world and Induced by a member from a family of corona viruses named (SARS COV-2) that has appeared in Wuhan in 2019 and can lead to sever acute respiratory syndrome with grave complication and even death of the infected person. The detection of persons infected with the virus is most important, as the virus as it can be easily transmitted from one to another and the person infected with the virus will also not know that he is infected until he has a number of symptoms. The detection of the virus is performed in this paper using deep learning as part of monitoring this outbreak, researchers began using software computing techniques to diagnose cases using chest CT-Scan images and X-Ray for the lungs, scan body temperatures and classify the severity of the disease. The research objective is to detect three classes: Covid-19 positive, Normal, and eumonia based on both X- Ray and CT-images. The importance of this research is to support the medical staff in Mosul city, in particular in the case of a heavy workload. The detection technique begins with some pre-processing tools for image processing to strengthen contrast, and then deep learning as the Convolution Neural Network (CNN) is used for detection. CNN will be based on the public dataset of COVID19 for training and forecasting other cases for the future. The programming language used in this paper is Matlab and the results of this study indicate that the best accuracy is obtained from the model with 99.55 %, 99.09 % sensitivity and 99.48 % precision of chest CT, but when the X-ray dataset is used The proposed model has achieved a classification accuracy of 85.58 %, 83.47% sensitivity and 87.33% precision.