5.3.3  Novel approaches: Artificial intelligence‐based technologies
Depp learning, as a novel AI‐based modality might be able to analyze radiographic features of COVID‐19, and help clinicians provide an accurate clinical diagnosis based on a precedented pattern (Wang, Kang, et al., 2020). As part of recent advancements, Convolutional Neural Network (CNN), a class of deep neural networks, has been shown to be capable of medical image analysis. To this date, CNN has been successfully employed in investigations on the nature of pulmonary nodules reported in CT images, diagnosis of pneumonia in children based on CXR, and image recognition in cystoscopy videos (Choe et al., 2019; Kermany et al., 2018; Negassi, Suarez‐Ibarrola, Hein, Miernik, & Reiterer, 2020; Wang et al., 2018).
The 21st century has seen many AI‐based models to be incorporated in several scientific fields, particularly imaging studies. Diagnostic AI‐based models might actually be a forward leap in tasks that simply cannot be handled by manpower, especially risk prioritization, that can greatly help improve patient turnaround time. Given the shortage of human resources and inadequate number of hospital beds in a country like China, AI‐based models for analysis of CXR and CT scans can be useful in ruling out irrelevant cases, and resource‐wise admission of patients to the hospitals (Kim, 2020).