A MobileNet-V2 COVID-19: Multi-class Classification of the COVID-19 by Using CT/CXR Images

COVID-19 is one of the most dangerous virus that has been separated among the entire world. At the beginning stage of COVID-19 virus, the RT-PCR is the only testing method to detect the virus. Later, the medical professions analyze the different medical scanning approaches for the detecting of COVID-19. The computer tomography (CT) and chest X-ray (CXR) images are well-suited for detecting the virus. In image processing algorithms, there is lots of deep learning (DL) algorithms are employed for identifying the diseases which are affected in the human body. Hence, the paper presents the deep learning approach of COVID-19 detection by using the CT/CXR medical images. Here, the pre-trained MobileNetV2 is fully loaded with training dataset of COVID-19 images. Initially, the testing medical images are preprocessed by DnCNN algorithm to get the residual image of the corresponding medical image and forwarded to the feature extraction unit, and finally, the classifier finds the COVID-19, non-COVID-19, and pneumonia from the testing dataset.