Application of Statistical Machine Learning and Deep Learning in Diagnosis of COVID-۱۹ through CT Images

Coronavirus disease, first detected in late ۲۰۱۹ (COVID-۱۹), has spread fast throughout the world, leading to high mortality. Chest CT scans have been reported to have sensitivity values of almost ۱۰۰%. The application of statistical machine learning and deep learning methods on CT images has facilitated the accurate diagnosis of COVID-۱۹. Statistical methods can be helpful to improve diagnosis of machine learning by image processing. The high incident rate of coronavirus defection in lungs and the late diagnosis show that this automated system can be conducive in the early scan stages. In this paper we use normalization of segmented lung images and help the convolutional neural network (CNN) model by clustering the lungs using k-means method. An exam of lung CT scan consists of a long series of images, and this system can analyze these images fast and reduce the risk of human error. This system can be used as the first step of a diagnosis; the marked cases can be passed for medical analysis for further studies and confirmation. A real sample of ۱۴۴ patients given the radiology center of Afshar hospital in Dezfool/Iran were analyzed in our study. We observed that our proposed method outperforms in classification of defected and no defected lungs by CT images

کلیدواژه ها:

Coronavirus ، Machine learning ، Deep learning ، Lung diagnosis ، CNN ، Clustering ، Segmentation