Artificial Intelligence Based COVID-19 Detection using Medical Imaging Methods: A Review

Abstract

The novel coronavirus disease (COVID-19), hereafter referred to as COVID-19, originated in the Wuhan district of China in December 2019. COVID-19 is a lower-respiratory-tract infection, and it can cause severe breathing problems and pneumonia. Symptoms of COVID-19 are similar to the other two most important viruses, namely Severe Acute Respiratory Syndrome (SARS) and Middle East Respiratory Syndrome (MERS). Currently, COVID-19 has been conclusively diagnosed through molecular tests ((polymerase chain reaction (PCR) and real-time reverse transcription-polymerase chain reaction test (RT-PCR)) with a high success rate. These tests are highly time-consuming, and may give false-positive results in COVID-19 detection. Moreover, they are invasive and require more facilities to perform molecular or rapid antigen tests (RAT). They require a well-trained human resource (epidemiologist or virologist). To circumvent these issues, researchers employed artificial intelligence (AI) methods to develop an intelligent clinical-diagnosis system for speeding up the early detection of COVID-19. Possibly, imaging could aid in screening or accelerate the speed of diagnosis, particularly with shortages of RT-PCR. One of the simple and non-invasive methods of COVID-19 detection can be performed through medical imaging methods such as chest X-Ray and chest CT scan images. These imaging devices are affordable, robust, non-invasive, and give accurate results as compared to RT-PCT in a short period. However, the resolution of X-Ray images is weak, and it makes it challenging to identify the most sensible feature of COVID-19 in X-ray images, such as ground-glass opacities (GGO). Hence, most of the researchers use chest CT scan images to effectively identify the GGO and multiple GGO present in the patients through high-resolution images. Hence, this chapter aims to review the different types of AI-based COVID-19 detection system using medical images such as Chest CT-scan and Chest X-ray. AI-based systems are beneficial to the doctors to diagnose COVID-19 patients, to decide what to do next depending on the output of the algorithm. It could also automate the diagnosis/prognosis of COVID-19 patients, to distinguish the COVID-19 patients to either mild or severe or non-severe COVID-19 and play as an alternative tool for RT-PCR test to identify COVID-19.