Prepoznavanje tumora mozga na slikama dobivenim magnetskom rezonancijom

undergraduate thesis

• Title (english): Using MR images for brain tumor detection
• Author: Ivan Mihaljević
• Mentor: Mirjana Domazet-Lošo (mentor)
• Committee member: Mirjana Domazet-Lošo (predsjednik povjerenstva)
• Committee member: Mario Brčić (član povjerenstva)
• Committee member: Mihael Kovač (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-12, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Strojno učenje nalazi sve veću primjenu u medicinskoj dijagnostici, a u ovom radu
korišteno je za utvrđivanje postojanja i raspoznavanje vrste tumora mozga. Dobrom
raspodjelom podataka kroz procese učenja, provjere i testiranja te pametnim odabirom
algoritama izgradnje moguće je napraviti kvalitetan model koji olakšava i ubrzava
dijagnozu bolesti. Snimke magnetske rezonance klasificirane su u četiri kategorije:
gliom, meningeom, tumor hipofize i snimka bez tumora. Jednostavnim metodama
obrade slike pokušalo se doći do boljih rezultata klasifikacije nego na početnom skupu
te je u tu svrhu napravljeno više modela koji su trenirani i na početnim i na uređenim
podacima. Modeli su napravljeni i trenirani pomoću Pythonove knjižice TensorFlow.
Rezultati ukazuju na vrlo dobru sposobnost dijagnostike razvijenih modela s obzirom
na njihovu relativnu jednostavnost i potencijal da uz dodatna istraživanja pronađu svoju
kliničku primjenu.

Abstract (english)

Machine learning is increasingly used in medical diagnostics. In this thesis it was
used to determine the existence and distinguish the type of brain tumor. Good data distribution
for learning, validation and testing processes and smart selection of construction
algorithms enables creation of a high-quality model that facilitates and accelerates
disease diagnosis. Magnetic resonance images are classified into four categories: glioma,
meningioma, pituitary tumor, and tumor-free images. Simple methods of image
processing were used to achieve better classification results than on the initial set, and
for this purpose several models were created and trained on both initial and processed
data. The models were built and trained using Python’s TensorFlow library. The results
indicate a very good diagnostic capability of the developed models considering
their relative simplicity and a potential for their clinical application.

• Keywords: tumor mozga; strojno učenje; umjetne neuronske mreže; konvolucijske neuronske mreže; klasifikacija; TensorFlow; Keras; analiza slike; obrada slike
• Keywords (english): brain tumor; machine learning; artificial neural networks; convolutional neural networks; classification; TensorFlow; Keras; image analysis; image processing
• Language: croatian
• URN:NBN: urn:nbn:hr:168:534200
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-10-11 09:49:22