15 Deep learning for analysis of COVID-19 electronic health records

The global pandemic coronavirus disease (COVID-19) is the reason behind all the chaos happening in medical world right now. There is a sudden surge of need for medical facilities around the globe, and prompted medical staff along with patients and families has very less time and information to take proper decisions, it is no surprise that hospitals are using computer vision, and artificial intelligence and machine learning techniques to assist with its functions. Healthcare is different from all other industries. It is a sector of high priority and people want the highest level of services, care without worrying about cost. But it is not possible to have medical experts mainly to interpret the medical data in remote areas of the country. This unprecedented crisis of COVID pandemic taught us how digital integration in healthcare is very important. It is very difficult to interpret the chest X-ray usually required to detect the level of infection in a COVID-19 patient, due to subjectivity and complexity of the images. Deep learning algorithms, such as convolutional neural networks (CNNs), are used for the automation of medical image analysis. This chapter provides study about deep learning architecture and its optimization used for medical image classification. In the pandemic of COVID-19, we can manage and improve patient care using deep learning algorithms especially in covert areas where medical experts cannot be physically present to analyze the images. In this chapter, we have discussed one such technique: image classification and applied on the dataset of the chest X-ray of patients to identify the infection of coronavirus. We will use our first CNN made using Keras and tensor flow on an open-source dataset of chest X-ray for identifying COVID-19 infection, and after building the model we will test our model using test data and try to achieve high accuracy.