Identification of Severity of Infection for COVID-19 Affected Lungs Images using Elephant Swarm Water Search Algorithm. (4th May 2022)

• Record Type: Journal Article
• Authors: Mandal, Sudip
• Is Part Of: International journal of modelling & simulation. Volume 42:Number 3(2022)
• Page Start: 518
• Page End: 532
• Publication Date: 2022-05-04
• Subjects: COVID-19 -- corona virus -- lung CT scan -- multilevel thresholding -- Elephant Swarm Water Search Algorithm -- optimization -- image processing -- Kapur's and Otsu's method; Mathematical models -- Periodicals; Simulation methods -- Periodicals; Mathematical models; Simulation methods; Periodicals; 003.3
• DOI: 10.1080/02286203.2021.1934797
• Languages: English
• ISSNs: 0228-6203

Abstract

ABSTRACT: Due to the outbreak of the pandemic COVID-19 or 'novel corona virus disease', the world is facing a global emergency. In case of severe infection, lungs are affected by COVID-19 significantly, whichmay lead to the death of the patient. In this paper, an automated image-assisted system based on artificial intelligence is proposed to extract infected sections from lung CT scan images that are caused due to COVID-19. Multilevel thresholding is a typical example of maximization problem of optimization to identify the threshold(s) levels for image segmentation. In this paper, Elephant Swarm Water Search Algorithm (ESWSA) has been used for multilevel thresholding based on Otsu's and Kapur's method and further analysis of lungs images. It has been observed from the obtained simulated results that ESWSA performs better than other state-of-the-art optimization techniques for multilevel thresholding. Moreover, location of infection and severity of infection has also been extracted from the pixel ratio between the infection and lung sections from the lung CT scan images. It is expected that the proposed methodology will support the doctors as it will reduce diagnostic burden and respective treatment process can be planned faster as the severity of infection by COVID-19 is found by this automated methodology.