Multiobjective Genetic Algorithm and Convolutional Neural Network Based COVID-19 Identification in Chest X-Ray Images. (25th February 2021)

• Record Type: Journal Article
• Authors: Shukla, Prashant Kumar; Sandhu, Jasminder Kaur; Ahirwar, Anamika; Ghai, Deepika; Maheshwary, Priti; Shukla, Piyush Kumar
• Other Names: Kaur Manjit Academic Editor.
• Is Part Of: Mathematical problems in engineering. Volume 2021(2021)
• Publication Date: 2021-02-25
• Subjects: Engineering mathematics -- Periodicals; 510.2462
• DOI: 10.1155/2021/7804540
• Languages: English
• ISSNs: 1024-123X

Abstract

Abstract : COVID-19 is a new disease, caused by the novel coronavirus SARS-CoV-2, that was firstly delineated in humans in 2019. Coronaviruses cause a range of illness in patients varying from common cold to advanced respiratory syndromes such as Severe Acute Respiratory Syndrome (SARS-CoV) and Middle East Respiratory Syndrome (MERS-CoV). The SARS-CoV-2 outbreak has resulted in a global pandemic, and its transmission is increasing at a rapid rate. Diagnostic testing and approaches provide a valuable tool for doctors and support them with the screening process. Automatic COVID-19 identification in chest X-ray images can be useful to test for COVID-19 infection at a good speed. Therefore, in this paper, a framework is designed by using Convolutional Neural Networks (CNN) to diagnose COVID-19 patients using chest X-ray images. A pretrained GoogLeNet is utilized for implementing the transfer learning (i.e., by replacing some sets of final network CNN layers). 20-fold cross-validation is considered to overcome the overfitting quandary. Finally, the multiobjective genetic algorithm is considered to tune the hyperparameters of the proposed COVID-19 identification in chest X-ray images. Extensive experiments show that the proposed COVID-19 identification model obtains remarkably better results and may be utilized for real-time testing of patients.