Ahmed Ossama Ibrahim, Marwan Hashem, Mostafa Ahmed, Omar Ahmed Salah
Supervised by: Dr. Alaa Hamdy, Eng. Verina Adel
November 8, 2021
The major goal of this project is to create an intelligent system that uses image analysis and machine learning algorithms to detect and classify various liver illnesses. This research will use Ultrasound (US) imaging, which has been determined to be one of the most prevalent and affordable types of medical imaging in Egypt , which is to take a non-invasive approach in the identification of liver illnesses. However, because of the low cost, the photos may be lacking in quality or detail. As a result, we will utilise a feature-based strategy to correctly categorise the US photos, and we will choose the most effi cient classifier (such as KNN, ANN, or SVM) to employ in the feature extraction phase after conducting extensive study.
Infection with the hepatitis C virus (HCV) is a major public health issue in Egypt. Egypt may have the highest infection rate of HCV in the world, with a frequency of 14.7 percent among those aged 15 to 59. HCV is one of the leading causes of cirrhosis and chronic liver disease. Pathology and histology have shown that the severity of liver disorders is linked to the advancement of liver fibrosis. As a result, most cases of liver disease can be divided into three categories: hepatitis, steatosis, and cirrhosis.
• Egypt has been endemic with cirrhosis and hepatitis C for decades, making it one of the world’s countries with the highest incidence of liver disease . Health-care resource constraints exacerbate the problem by enabling disease-specific consequences to emerge in the absence of effective therapies in the past. Hepatitis C, which is still a major health problem in Egypt despite the availability of extremely efficient interferon-free antiviral regimens, is a case in point. This is mostly because to the huge number of infected patients (estimated to be over 6 million), re-infection, and the existence of comorbidities that require additional therapy.
• The need to address these important health issues in Egypt prompted a great deal of hepatology research. Many scientific publications about liver illnesses in Egypt have been published by Egyptian and non-Egyptian researchers. Needle biopsies have historically, and still remain, the go-to method of accurately identifying liver disease , but they are still an invasive and dangerous procedure. The alternative non-invasive techniques that have been employed are medical imaging such as Ultrasound, CT-scan and MRI, which help medical professionals identify these diseases without the need for invasive procedures.
• The MRI continues to be a powerful imaging tool that can be used to accurately diagnose diffuse liver disease, but the accuracy and quality of the imaging results in a very expensive process. This leads to many people in rural areas avoiding processes like MRI scans as they can rarely afford such a luxury. When it comes to cheap imaging techniques that still produce quality results, US imaging reigns supreme across the board. 1.2.2 Business
• A program that shall accurately diagnose a patient using a simple US image in a matter of minutes provides doctors with a very useful tool at their disposal. This program can be used in hospitals to help medical professionals know where to start a diagnosis or as a reaffirmation of the diagnosis itself or even to confirm it. It can also be used to teach medical students how they can get a quick overview of what to look for in an ultrasound image.
1.3 Problem Statement
Diffuse liver diseases, Hepatosteatosis & Liver Cirrhosis, have been around for centuries, and have been two of the most fatal diseases in Egypt. Therefore, the best course of action is to detect these symptoms at the earliest convenience. Through Ultrasound (US) imaging, the proposed system will be able to classify US images into three distinct classes; Healthy Liver, Steatotic Liver, Cirrhotic Liver, and classify the diseases into stages; mild, moderate and severe. However, US imaging itself poses a problem of producing low-quality images which makes the tasks of feature selection and extraction more difficult.