Classification and Identification of Infectious COVID-19 Virus Using Deep Learning and Machine Learning Techniques: A Comprehensive Analysis

Early detection and diagnosis of the infectious COVID-19 virus is a challenge due to the shortage of kits and rapid growth in every part of the world. The rising scope of Machine Learning (ML) and Deep Learning (DL) techniques can ease the problem of detecting the virus in the healthcare area and can also solve the problem. The paper’s main objective is to find out some of the practiced techniques and methodologies utilizing Chest X-rays and CT scans in this area to classify, detect, and diagnose the infectious COVID-19 virus and to find out the advantages, disadvantages, and limitations of the models. DL and ML are growing techniques, especially in imaging that could rapidly help diagnose COVID-19. The selection of the articles has been performed through the PRISMA guideline. The presented findings include country-wise publications, frequency of articles, data collected from various databases, analysis of the datasets, and the AI techniques included in the diagnosis. From the study, it is found that there are a lot of ML and DL techniques available to detect and classify COVID-19 with an acceptable performance and accuracy rate of 75–96% when validated with Chest X-rays, and the performance accuracy is lower in the case of Computed Tomography scans. However, very few techniques have shown sensitivity and specificity in the diagnosis of the infectious virus at a low rate. Even the time complexity and the intensity of the disease have also not been discussed in any of the articles, which is an important aspect. The work is concluded with proposals for future recommendations for different frameworks (DL/ML) that are robust, efficient, and sensible to detect the infectious COVID-19 virus rapidly and within a low period.