Evaluation of deep learning models for identification and prediction of new strains of COVID-19

COVID-19 virus has caused an unembellished threat to mankind and industry over the last 2 years. Currently, when the entire world is reeling from the effects of the second wave of the pandemic, the third wave is in its beginning and is expected to be more deadly than earlier waves. Ever since the first Covid case was detected, the virus has undergone several mutations, with the delta plus variant, the delta variant, and the lambda variant being a few of the worst COVID-19 variants detected so far. During the last few months, there has been an exponential rise in the number of Omicron cases all over the world. Proper identification and detection of these new variants are vital in the current times. In this chapter, various deep learning (DL) and transfer learning models like convolutional neural network (CNN), Resnet50, VGGNet, and InceptionV3 have been explored that can help to forecast the spread and further mutations of the virus. The objective of the chapter is to study the performance details of all those algorithms comprehensively for prediction of Covid cases. As deep learning models are powerful in extracting complex features automatically without human effort, these models can formulate prevention strategies for the new strains of Covid.