College of Engineering

A Master Thesis was discussed in Department of Computer Engineering at College of Engineering / University of Mosul, entitled “A Deep Learning System for Automatic COVID-19 Detection” on Monday, Sep. 6, 2021, submitted by postgraduate student (Saja Waled Mahmood).The thesis dealt with an automated deep training by ten models for detecting COVID-19 in X-ray/CT scan is proposed. Two deep Pre-processing of the two data sets are performed X-ray / CT to standardize the data in terms of size and appearance, in addition to enhancing them by Zeke and Gaussian Filters. A Grab cut segmentation is proposed to separate the injury portion in the CT images. All these steps are led to an increase the accuracy of the diagnosis to 99.94 in the X-ray images using Xception model, and 99.95 in the CT scan data utilizing InceptionV3 model. The proposed system can be linked with X-ray/CT imaging devices to automatically deliver the results of the examination to the patient. These are sent to the patient’s phone only if the result is negative, and to three parties if it is positive, in addition to the patient’s phone, email to the health center and to the community via WhatsApp.Cough is a third deep training data in the proposed system, that can be recorded by personal mobile and sent through proposed CXC web site and received the testing result via phone at same previous procedures.The results showed that the proposed system can be modeled to utilized in any health center with any other diseases by change training data set only.