Multi-classification model for covid-19 prediction using imbalanced x-ray dataset based on transfer learning and class weighting-smote method

OLA M. EL ZEIN; MONA M. SOLIMAN; A. K. ELKHOLY; NEVEEN I. GHALI

No Detail Found!

Multi-classification model for covid-19 prediction using imbalanced x-ray dataset based on transfer learning and class weighting-smote method

Author(s)

OLA M. EL ZEIN Faculty of Science, Al-Azhar University (Girls branch), Cairo, Egypt.

MONA M. SOLIMAN Faculty of Computers and Artificial Intelligence, Cairo University, Cairo, Egypt.

A. K. ELKHOLY Faculty of Science, Al-Azhar University (Girls branch), Cairo, Egypt.

NEVEEN I. GHALI Faculty of Computers and Information Technology, Future University in Egypt, Cairo, Egypt.

Abstract

This paper proposes a deep neural networks model to predict COVID-19 patients automatically based on chest X-ray images. The model is trained using imbalance dataset with a new hybrid balancing technique proposed to solve this problem. The Deep Convolutional Neural VGG-16 is trained and utilized to extract features from a given chest X-ray image after some preprocessing steps. To overcome the data imbalance issue, a new hybrid Class Weights-SMOTE is applied to the extracted feature vector and compared with traditional balancing techniques. The feature vector is then classified utilizing a Fine-tuning VGG-16. The model provides a multi-classification for the input x-ray images into COVID-19, Normal, and Pneumonia. Comparison with existing methods shows that the proposed model achieves a superior classification accuracy and outperforms all other models, providing 98% accurate prediction and improving the model's performance on minority-class samples to achieve high accuracy 100%. The findings of this study could be useful for diagnosing COVID-19 from chest X-ray images.

Publication Details

Page(s) 1354-1368

DOI DOI not available

Published Journal: Journal of Theoretical and Applied Information Technology, Volume: 100, Issue: 5, Year: 2022

Request the full text from PASTIC

Keywords

Coronavirus, Class Weight, Gaussian Filter, Imbalanced data, SMOTE

Citations

Downloads

Views