Title: Design of optimised lung lobe segmentation and deep learning model for effective COVID-19 prediction

Authors: Anandbabu Gopatoti; P. Vijayalakshmi

Addresses: Department of Electronics and Communication Engineering, Hindusthan College of Engineering and Technology, Coimbatore, Tamil Nadu, India ' Department of Electronics and Communication Engineering, Hindusthan College of Engineering and Technology, Coimbatore, Tamil Nadu, India

Abstract: The Taylor improved invasive weed sea lion optimisation (Taylor IWS) algorithm is introduced in this paper for the identification of coronavirus (COVID-19) by chest X-radiation (X-ray) images. The developed Taylor IWS technique is the integration of the Taylor series, sea lion optimisation (SLnO) and improved invasive weed optimisation (IIWO) algorithms. The input is pre-processed using an adaptive thresholding model. Then, lung lobe segmentation is done using U-Net and SegNet models, which are trained using the developed Taylor IWS, and trained output is fused using cosine similarity. After that, the multi local texture feature (MLTF) is extracted. Finally, the COVID-19 detection is determined by Taylor IWS trained deep convolutional neural network (DCNN). The developed model attains maximum accuracy, true positive rate (TPR), and true negative rate (TNR) of 0.9378, 0.9552, and 0.9094.

Keywords: deep convolutional neural network; DCNN; sea lion optimisation; SLnO; multi-local texture features; U-Net; SegNet.

DOI: 10.1504/IJBIC.2023.133507

International Journal of Bio-Inspired Computation, 2023 Vol.22 No.1, pp.16 - 27

Accepted: 21 Apr 2023
Published online: 18 Sep 2023 *

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