ABSTRACT
The WHO has declared Human Coronavirus (HCoV) ongoing outbreak to be a global public health emergency. Corona virus (HCoV) was reported two months ago in Wuhan, China. Health care systems over the world get into a chaotic mode due to limited capacity and a hectic increase of suspected coronavirus cases. The one thing that everybody is trying to do is to reduce the effect of cause created for a patient. This study will show how Machine Learning technique can be used for classifying the infected and healthy lung using the nano scaling imaging technique of computed tomography (CT) lung scans. Pre-processing is used to reduce the effect of intensity variations and for noise removal between CT slices. Then thresholding and other morphological operation is used to separately isolate the background of the CT lung scan. Each dataset that we take undergoes a texture-based feature extraction method in which it uses GLCM along with a wrapper method for optimization. The obtained features are classified using a Deep convolutional neural network, which will classify in several layers. By giving our input of scan images it will train in an efficient manner and gives us an accuracy of 99%.