Data Science and Intelligent Computing Techniques
Predicting COVID-19 Using Deep Learning: A Comparative Study
Authors: Vaishnavi Jariwala, Suraj Patil and Dhananjay Joshi
Publishing Date: 27-05-2023
ISBN: 978-81-955020-2-8
Abstract
In 2020, WHO declared COVID-19, an infectious disease caused by SARS-CoV-2 virus as a global pandemic. With the increase in COVID-19 cases worldwide, it becomes very crucial to control and manage the spread of the virus. The disruption caused by the virus has impacted the lives of many people and affected various sectors beyond repair. Applications of Deep Learning Machine Learning can be used to detect various diseases including COVID-19. This study reviews recent studies on various machine learning applications for the detection of COVID-19 via CT and CXR images. This study consists of extensive research of 60 articles. Various aspects which include dataset preparation, feature extraction, classification algorithms, and model evaluation have been discussed in this study. Various ImageNet algorithms such as ResNet, VGG, AlexNet are discussed in literature review of this study. Some of the studies used techniques such as transfer learning and Support Vector Machines for classification purposes. It was found that Convolution Neural Networks (CNN) and Transfer Learning were the most used techniques. Many studies describe how overfitting and gradient vanishing problems can be avoided in a model. For model evaluation, various metrics such as accuracy, recall, specificity, precision, F1-Score, ROC curve, and cross-validation are used by many studies. All these studies conclude that the application of Deep Learning for early detection of COVID-19 can be a significant tool for the healthcare domain. Moreover, these techniques can save the time of radiologists in detecting the disease and necessary measures can be taken for the people diagnosed with COVID-19 quickly and effectively. Therefore, applications of these methods may prove effective in detecting the diseases in early stages thus saving time, cost, and lives, hence proving beneficial to mankind.
Keywords
Automated detection, Convolutional Neural Networks, COVID-19, Deep Learning, Diagnosis, Machine Learning, Medical imaging.
Cite as
Vaishnavi Jariwala, Suraj Patil and Dhananjay Joshi, "Predicting COVID-19 Using Deep Learning: A Comparative Study", In: Satyasai Jagannath Nanda and Rajendra Prasad Yadav (eds), Data Science and Intelligent Computing Techniques, SCRS, India, 2023, pp. 209-225. https://doi.org/10.56155/978-81-955020-2-8-18