Deep Learning Study of Pandemic Contact Tracing

Abstract

The outbreaks of infectious diseases exemplified by the COVID-19 pandemic, have left the global health community with the continuous search for novel methods that can help in emergencies. This is because traditional pandemic contact tracing methods have failed in enhancing accuracy, efficiency, and adaptability within pandemic scenarios. In this study, deep learning study of pandemic contact tracing was investigated. Deep learning models were developed using neural networks, namely, Convolutional Neural Networks (CNNs). Networks dataset of images which capture individuals and their interactions were used as the CNN input data and to analyze the transformed data. Results reveal an accuracy of 97% from the CNN. Finding further reveal that the CNN model in the present study compares favourably with the results of other traditional pandemic contact tracing model such as MobileNetV2, VGG16, and InceptionV3. Based on these results, it was therefore recommended among others, that the CNN model holds significant promise for the future of pandemic contact tracing due its scalability and adaptability in emergency situations. Overall, the adoption of the CNN by health workers, could lead to superior accuracy and efficiency in identifying potential contacts and exposures, a critical aspect in curbing the spread of diseases during pandemics.