Computer Vision-Based Prognostic Modelling of COVID-19 from Medical Imaging

ABSTRACT

COVID19 is a respiratory illness that is extremely infectious and is spreading at an alarming rate at the moment. Chest radiography images play an important part in the automated diagnosis of COVID19, which is accomplished via the use of several machine learning approaches. This chapter examines prognostic models for COVID-19 patients’ survival prediction based on clinical data and lung/lesion radiometric characteristics retrieved from chest imaging. While it seems that there are various early indicators of prognosis, we will discuss prognostic models or scoring systems that are useful exclusively to individuals who have received confirmation of their cancer diagnosis. A summary of some of the research work and strategies based on machine learning and computer vision that have been applied for the identification of COVID19 have been presented in this chapter. Some strategies based on pre-processing, segmentation, handmade features, deep features, and classification have been discussed, as well as some other techniques. Apart from that, a few relevant datasets have been provided, along with a few research gaps and challenges in the respective sector that have been identified, all of which will be useful for future study efforts. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.