PANDIT: An AI Twin-Based Radiography Image-Assisted nCOVID-19 Identification and Isolation

ABSTRACT

The unprecedented rise and spread of the pandemic in form of nCOVID-19 has really raised high concerns in the socioeconomic front. The usual diagnosis is made by an RT-PCR test, which is highly specific can incorrectly identify some nCOVID-19 individuals to cause a serious compromise in overall accuracy. Since the drug application in its full swing is still some months away, hence, the need of the hour is to find a more accurate technique which can be used by health care centers having basic point of care facilities. The increase in the number of cases in India and lack of test kits in some of the less known diagnostic centers has added more concerns to the increasing problems. Additionally, the test kits incur a significant cost making it less affordable to some of the diagnostic centers. Hence, this research group in this article has proposed an algorithm centered around the concept of Internet of Things, a dual deep learning based algorithm, and collating the decision by a strong decision fusion technique. The objective of the algorithm is to detect and isolate the nCOVID-19 subjects in a cost-effective way to keep a check on the spread. This pandemic detection and isolation technique (PANDIT) is based on two different radiography image technology and uses a state-of-the-art deep learning algorithm for the purpose. The radiography technique has long been the most acceptable technique for cases related to pneumonia. The group has developed the algorithm based on X-ray and CT scan as its training data. The novelty of this paper is best described by a multi-fold methodology. Firstly, the significance of radiography imaging for detecting and identification of COVID-19 subjects. A simple connected value chain driven by Internet of Things (IoT) would enable the isolation process in an efficient and accelerated manner. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.