Prediction of Covid-19 disease with Resnet-101 deep learning architecture using Computerized Tomography images
Bekir Aksoy, Osamah Khaled Musleh SALMAN
Abstract
Many pandemics have caused the deaths of millions of people in world history from past to present. Therefore, the measures to be taken in the prevention of pandemics are of great importance. In addition to the precautions, it is very important to be able to diagnose the disease early. The most recent pandemic occurred in the world is the COVID-19 outbreak that emerged in China in late 2019. In this study, Computerized Tomography images of 746 patients taken from an open source (GitHub) website were used. Images were analyzed using the Resnet-101 model, which is one of the deep learning architectures. Classification process was carried out with the created Resnet-101 model. With the Resnet-101 model, individuals with Covid-19 disease were tried to be identified. The Resnet-101 model detected individuals with Covid-19 disease with an accuracy rate of 94.29%.