Predictive Analysis for the Detection of Covid-19 with Chest X-Ray Images Using Convolutional Neural Network

Abstract

Artificial Intelligence comes up with a lot of ease and advancements in almost all sectors of living. No one of us can deny its contributions in the medical field. Disease detection is one of the greatest achievements of Machine Learning. During the pandemic of COVID-19, medical emergency and less of experts has affected the health sector a lot. Detection of Covid-19 has become much more important than its cure to protect others from the virus. Detection of Covid-19 with our model is much easier through the x-ray images. The model using Convolutional Neural Network has trained on our self-made algorithm which was named to be Lungs X Ray Neural Networks (LxN) providing much more accurate than any other model available. It can process multiple datasets in a batch and our model is generalized very well with an accuracy of 98.8 % on validation and 98.0% on test set. The dataset for solving this problem was obtained from the open-source ieee8023 GitHub Repository, constantly updating with the images around the globe, containing a combo of both corona and non-corona cases. The result obtained from the model is y ? {1, 0} indicating Corona presence or absence respectively. Lungs X Ray Neural Networks (LxN) model is not only specialized on corona data but can be fine-tuned on any other lung images like pneumonia. Thus LxN has significant role in the research area of AI in case of medicine.