ABSTRACT
Machine Learning and Deep learning are being leveraged by almost every field across the globe. They have now become ubiquitous tools for research in medical and health sciences. A highly rapid response is critically essential in the treatment of acute neurological illness. Henceforth, the techniques and tools that reduce the diagnosis time lead to improved treatment for patients. The data sets of medical imaging are imbalanced as pathological findings are very rare which aims at introducing new challenges. To improve the performance of learning systems, finding patterns among multimodal data is essential. In general, the data skewed in terms of ethnicities, age groups and other characteristics could create models that are not well-defined to access a variety of patients. Accessing high quality data to train the models is really challenging. Deep learning presents a promising pathway to break through the existing barriers. This chapter focuses on frontiers of Deep Learning and Radiography to diagnose COVID-19. The motive is to accurately screen the patients suffering from COVID-19 against those who suffer from Ground Glass Opacity and Viral Pneumonia which have a similar effect on human lungs as that of COVID-19. The comparative analysis of deep Convolutional Neural Networks namely InceptionResNetV2, ResNet152V2, Xception, DenseNet201 and Mobilenet based on their test Accuracy, Precision, F1 score and Recall as the evaluation metrics are done to diagnose COVID-19.
