Prediction of COVID-19 Patients from X-Rays using Ensemble Deep Transfer Learning Techniques

Abstract Deep learning is very effectively used in the medical field to predict diseases such as pneumonia classification , cancer, and so on. Convolutional Neural Network (CNN) is used to classify the chest X-rays of humans in normal and COVID-19 infected cases. To build a model, pre-trained models and transfer learning is used. In this paper, four pre-trained models VGG16, VGG19, ResNet50, and InceptionV3 are ensemble and build a new ensemble model. To train and test a model, a dataset of as many as 720 X-ray images has been used. Less number of images makes us apply image augmentation. The ensemble models such as VGG19 and ResNet50, VGG16 and ResNet50, VGG16 and InceptionV3, and VGG19 and InceptionV3 yield 96.53%, 98.61%, 99.31%, and 100% accuracy respectively.