University of Patras
Συστήματα ανίχνευσης χαρακτηριστικών των ασθενειών με τεχνικές μηχανικής μάθησης
Abstract
dc:descriptionSystems to detect traits of health diseases using machine learning techniques One of the fastest growing areas of computer science is the development of intelligent systems that support user decisions and provide a wide range of services. Their intelligence is based on monitoring and decoding the user's data and needs and is achieved by applying machine learning techniques. Especially image processing, object detection and the extraction of disease characteristics/properties from all kinds of input data are the most important applications of machine learning techniques in Medicine. Earlier, before the era of Machine Learning, the process of feature extraction was limited. Neural networks constitute the methods to achieve both feature extraction and feature evaluation, thereby reaching decisions by filtering out irrelevant extracted information. The data are processed and represented through mathematical processes and this representation leads to the extraction of innumerable properties–characteristics extracted from the input images. The purpose of this thesis is to explore the benefits of machine learning techniques in medicine, such as image classification, image feature extraction, human organ segmentation, and disease severity and biomarker assessment. The interpretation of the results is done by both fields, medicine, and informatics. The aim is to present models for dynamic assessment of the severity of a disease and extraction of its important characteristics. This work concerns liver cancer, colitis, respiratory problems such as asthma. More specifically, these models could be used as diagnostic tools in real time and as auxiliary systems for making medical decisions in real conditions. The systems approach is applied starting with the input elements. These data are about the patient: mouse or human (characteristics) and the traits of disease: liver cancer, colitis, respiratory problem like asthma. We continue with the intelligent - algorithmic, statistical, etc. - processing of these data (a) to assess the current state and severity of the disease, patient characteristics and conditions and (b) to understand the situation and passively or actively support the doctor or health researcher, with the relevant diagnosis or prognosis. The scanning of human organs and its results in the form of Magnetic Resonance Imaging (MRI) or Computed Tomography (CT) plays an important role in the diagnosis and classification of lesions/diseases such as cancerous tumors, and/or in evaluating the degree of severity some disease. The aim of this thesis is to propose the best options for developing and implementing machine learning training models for information systems, mobile applications, and web apps to detect diseases and their characteristics, the type of experimental sample. Also, research focuses on detecting features of various diseases, suchas asthma or ulcerative colitis, through machine learning algorithms applied to MRI/CT or histological image data. The detection and analysis of disease characteristics is done without requiring prospective data from the respective application or system user. The research includes statistical analysis of all the training algorithms of the proposed systems, based on the calculation of values of a variety of metrics. It offers the possibility of updating ,monitoring, re-evaluating diseases such as cancer, colitis, asthma, etc., through an easy-to use mobile or web application. The goal of the research is the development of intelligent information systems that could significantly help in diagnosis and prognosis, immediately and accurately. The proposed methods could be used in real-world conditions by hospitals and health research laboratories.
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Giannou, Olympia
- Γιάννου, Ολυμπία
Subjects
dc:subject × 26- Μηχανική μάθηση
- Ιατρικά συστήματα
- Αρχιτεκτονικές λογισμικού
- Πρόγνωση
- Διάγνωση
- Συστήματα
- Έξυπνα συστήματα
- Ασθένειες
- Αυτοματισμός
- Machine learning
- Medical systems
- Software architecture
- Prognosis
- Diagnosis
- Systems
- Smart systems
- Diseases
- Automation
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Αυτοματισμοί και Συστήματα ελέγχου
- Τεχνολογία μέσων
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Automation and control systems
- Media Technology
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/53526