Study of X Ray Detection Using CNN in Machine Learning

ABSTRACT

The coronavirus spread that started in Wuhan, China and spread across the world, affecting the best of the healthcare systems from the Lombardy region of Italy to India, the US, and the UK, required accurate diagnosis. A rapid assessment to ascertain whether or not a patient has COVID-19 is required by frontline clinicians. In this paper, we propose to deduce the presence of COVID-19 using X-ray images of the lungs through feature extraction. A convolution network model is built for binary classification of images into corona positive and negative using the deep learning framework on Python, Keras. Various studies using different classifiers such as CART, XGB-L and XGB Tree were studied, which used machine learning for detection of COVID-19 and yielded a very accurate diagnosis. In this particular CNN model, Google Colab is used to execute the algorithm. The dataset is trained and the validation accuracy obtained is more than 96%. This is a very cost-effective way of using machine learning for the classification of infected and non-infected cases since working on Google Colab doesn't require enormous computational resources. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.