Project Home
PDF access
Authors
Noran Hany, Nura Mostafa, Nourhan Atef, Sara Mohamed
Publishing Date
December 2020
Abstract
The main idea of this project is to detect COVID-19 and pneumonia and discriminate between them. COVID-19 is an infectious disease that mainly affects lungs. Pneumonia is also an infection that causes inflammation in the air sacs in lungs. COVID-19 and pneumonia have similar symptoms in lungs as breathing problems that appear in CT as ground-glass opacity (GGO). For discrimination between them, we are using a dataset that consists of CT images of three classes (COVID-19, Pneumonia, Normal lungs), for determining the severity of each of them, we are using K-means clustering. And for COVID-19 detection we will use a dataset which contain 6 blood tests and PCR for COVID-19 and NON-COVID-19 patients. By using deep learning techniques, which will facilitate diagnosing these diseases efficiently and in a shorter time, it is expected to provide better results compared to other techniques. The Convolutional Neural Networks model (CNN) is proposed to be used in differentiating between these two diseases and get better results. It is considered a high sensitivity model to characterize and diagnose COVID-19 infections.
1.1 Purpose of this document
The purpose of this document is to determine the functional and non-functional requirements for our system that’s main target is to discriminate between COVID-19 and Pneumonia. Many doctors are facing difficulties diagnosing them because they both have similar symptoms, especially in the lungs. So, the system uses CT scans instead of PCR test. As the results of PCR test that is not accurate and reported to suffer from high false negative rates. Therefore, not only doctors but also patients can use our web application to know if they are suffering from COVID-19 or Pneumonia.
1.2 Scope of this document
The scope of this document is to specify and analyse the software requirements of our system. This document gives a detailed description of our application. It targets COVID-19 and Pneumonia patients which made a huge number around the whole world recently and the doctors that are related to this field. Users (patients and doctors) will find this application an easy way to diagnose the two diseases. Patients will use this system to decrease the probability of infection while going to the crowded hospitals and clinics and get faster results. Doctors will use this system to avoid the misdiagnosing of the two diseases as they have similar symptoms, and assist them in detecting easier and faster.
1.3 System Overview
The main goal of our system is to detect COVID-19 and Pneumonia diseases and discriminate between them in CT images of lungs by using By using deep learning techniques (CNN Model). And classify the severity of the two diseases in the patient’s lungs from CT. Also, detect COVID-19 through 6 blood tests which are labeled by PCR test. These blood tests might help to identify false-positive/negative PCR tests using machine learning algorithms.
1.4 System Scope
This system is a web application, that can be accessed by doctors and patients everywhere. It aims to detect both diseases (COVID-19& Pneumonia) that affect the lungs, discriminate between them from CT scan images of lungs and detect COVID-19 with the help of six blood tests. The blood test labeled by PCR test. Also the system determines the severity(moderate or severe) of these two diseases by clustering using Deep Learning algorithms. This system will achieve a high accuracy, by using CT instead of x-ray images which does not provide enough detailed description of the lungs compared to CT.