Title: Automated COVID-19 detection from chest X-ray and CT images using optimised hybrid classifier

Authors: Madhavi Bhongale; Pauroosh Kaushal; Renu Vyas

Addresses: School of Bioengineering Sciences and Research, MIT Art, Design and Technology University, Pune, India ' School of Bioengineering Sciences and Research, MIT Art, Design and Technology University, Pune, India ' School of Bioengineering Sciences and Research, MIT Art, Design and Technology University, Pune, India

Abstract: Amidst the global threat of infectious diseases, exemplified by COVID-19, conventional RT-PCR detection methods are time-consuming and potentially misleading. This study introduces an innovative approach, utilising CT and X-ray images as markers for efficient COVID-19 detection. An automatic assessment tool, integrating V-SLBT and GLCM features, optimises image texture analysis for precise classification by a deep belief network (DBN). Enhancing accuracy, a hybrid BWUCOA is integrated into DBN. The tool's workflow involves image preprocessing, optimal texture feature computation, and DBN-based classification. Validation with clinical data from 82 patients attests to a 98% accuracy. Comparative analysis reveals a 1.32% improvement for X-ray and a 2.38% enhancement for CT images over existing methods, underscoring the efficacy of V-SLBT and BWUCOA in refining the classifier's accuracy. This swift and cost-effective tool provides a precise diagnosis for COVID-19.

Keywords: COVID-19 detection; CT image; chest X-ray image; GLCM; SLBT feature; deep belief network; DBN; black widow updated coronavirus optimisation algorithm; BWUCOA.

DOI: 10.1504/IJBET.2024.140559

International Journal of Biomedical Engineering and Technology, 2024 Vol.45 No.4, pp.269 - 295

Received: 05 Jul 2023
Accepted: 03 Dec 2023
Published online: 23 Aug 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article