A Robust Technique for Detecting SARS-CoV-2 from X-Ray Image using 2D Convolutional Neural Network and Particle Swarm Optimization

Abstract

The coronavirus (COVID-19) that was unleashed in Wuhan, China, has recently traveled the world and the recent pandemic is rapidly spreading around the world through hands and breadth. Strategies to deal with this start with diagnosing those affected. A real-Time Reverse-Transcription-polymerase reaction is a very slow and unreliable means if diagnosing the disease thus a much quicker alternative is required. Initially, we would collect and arrange all our data through the methods of data synthesis. Another important factor is the testing and training of the proposed 2D Convolutional Neural Network (2D-CNN) model, dividing the data into three groups, and each group consists of different images for testing, training, and validation. The features are then selected using the Particle Swarm Optimization (PSO) method and smoothened using Principle Component Analysis (PCA). Data evaluation is done through classification using a support vector machine (SVM) and an array, which provides the results necessary to record and diagnose diseases based on a variety of data samples.