Download PDFOpen PDF in browser

DNN-GFE: A Deep Neural Network Model Combined with Global Feature Extractor for COVID-19 Diagnosis Based on CT Scan Images

EasyChair Preprint 6330, version 2

Versions: 12history
17 pagesDate: August 19, 2021

Abstract

Background:

COVID-19 is a viral disease that was first observed in Wuhan, China in December 2019. This virus can be influenced organs, especially the lungs. The testing kit is one of the most common diagnostics tools for the virus that clinical centers are facing a shortage of due to the increase in patients. Also, X-ray and CT scan images play a crucial role for diagnosing. To the best of our knowledge, artificial intelligence-based methods have been utilized for COVID-19 diagnosis on the images.

Materials and Methods:

Hence, a Deep Neural Network model combined with a Global Feature Extractor called DNN-GFE is proposed for COVID-19 diagnosis with 1252 sick and 1229 Healthy CT scan images. The DNN-GFE model is developed to improve as an accurate diagnostics for classifying Sick and Normal persons. Also, the other classification models such as Decision tree, Random forest, and Neural net had been generated. In this paper, image normalization is used to create good quality images and improve the diagnosis of COVID-19. Furthermore, the 10-fold cross-validation technique is utilized for partitioning the data into training, testing, and validation.

Results:

The experimental results of the DNN-GFE model were compared with three classification models for COVID-19 diagnosis regarding accuracy so that the DNN-GFE model has the most accuracy of 96.71%.

 

Conclusion:

The proposed model has the best performance regarding evaluation metrics over the existing models that can be utilized instead of a testing kit for COVID-19 diagnosis for clinicians.

Keyphrases: Artificial Intelligence, COVID-19 Diagnosis, CT scan images, Deep Neural Network, machine learning

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
@booklet{EasyChair:6330,
  author    = {Javad Hassannataj Joloudari and Faezeh Azizi and Issa Nodehi and Mohammad Ali Nematollahi and Fatemeh Kamrannejhad and Amir Mosavi and Edris Hassannatajjeloudari and Roohallah Alizadehsani},
  title     = {DNN-GFE: A Deep Neural Network Model Combined with Global Feature Extractor for COVID-19 Diagnosis Based on CT Scan Images},
  howpublished = {EasyChair Preprint 6330},
  year      = {EasyChair, 2021}}
Download PDFOpen PDF in browser