Predikcija starosti mozga novorođenčeta temeljem značajki iz strukturnih MR slika

undergraduate thesis

• Title (english): Neonatal brain age prediction based on features from the structural MR images
• Author: Mateja Vuradin
• Mentor: Jelena Božek (mentor)
• Committee member: Jelena Božek (predsjednik povjerenstva)
• Committee member: Filip Turčinović (član povjerenstva)
• Committee member: Ivana Žeger (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Brojni današnji problemi i zadaci vrlo su kompleksni, osobito za računala. Ovdje na snagu nastupa efikasnost modela dubokog učenja, tj. umjetnih neuronskih mreža. Njihov cilj je brzo nalaženje rješenja, ali da pritom ostanu generalizirane. U ovom radu istražila se njihova primjena na MR slike mozga novorođenčadi s ciljem predviđanja starosti. Ovo je ključno za razvoj medicine i rano otkrivanje bolesti djece. Koristila se metoda regresije, klasična metoda predviđanja kontinuiranih vrijednosti korištenjem nadziranog učenja. Istražene su metode i model iz FastMONAI knjižnice, a baza podataka uzeta je iz dHCP-a. Dobiveni rezultati na podacima za testiranje bili su zadovoljavajući s pogreškom manjom od 1,5 tjedana starosti.

Abstract (english)

Numerous today's problems and tasks are highly complex, especially for computers. This is where the efficiency of deep learning models, i.e., artificial neural networks, comes into play. Their goal is to quickly find solutions while remaining generalized. This study explored their application to MRI of newborn brains with the aim of predicting age. This is crucial for the development of medicine and early detection of childhood diseases. Regression, a classical method for predicting continuous values using supervised learning, was utilized in this work. Methods and models from the FastMONAI library were explored, and the dataset was obtained from dHCP. The obtained results on the test data were satisfactory, with an error of less than 1,5 weeks of age.

• Keywords: duboko učenje; konvolucijske neuronske mreže; magnetska rezonancija; T1 slike; T2 slike; FastMONAI
• Keywords (english): deep learning; convolutional neural networks; magnetic resonance imaging; T1 images; T2 images; FastMONAI
• Language: croatian
• URN:NBN: urn:nbn:hr:168:376507
• 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: 2023-07-13 10:41:34