Analiza stomatoloških rendgenskih snimaka za procjenu starosti pacijenta

master's thesis

• Title (english): Analysis of Dental X-ray Images for Estimating the Age of the Patient
• Author: Marin Drabić
• Mentor: Marko Subašić (mentor)
• Committee member: Marko Subašić (predsjednik povjerenstva)
• Committee member: Tomislav Pribanić (član povjerenstva)
• Committee member: Zvonko Kostanjčar (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-07-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj ovog diplomskog rada je istražiti razne algoritme za analizu informacija prisutnih u stomatološkim rendgenskim snimkama (ortopantomogramima) pomoću dubokog učenja. S naglaskom na implementaciji dubokih umjetnih neuronskih mreža koje su inspirirane biološkim neuronskim mrežama koje tvore mozak, te korištenju predtreniranih detektora značajki.

Ortopantomogram ili ortopan je slika kompletne gornje i donje čeljusti s okolnim anatomskim strukturama koji daje opću informaciju o kostima glave, sinusima, čeljustima i zubima. Njegova primjena je u tome što omogućuje procjenu stupnja dozrijevanja zuba, redoslijeda i brzine njihova nicanja te procjenu dentalne dobi pacijenta.

Primjenjivost se ispituje na procjeni dentalne dobi pacijenata iz oropantomograma, kao i iz parcijalnih rendgenskih snimaka ljudske čeljusti. Postignuti su rezultati procjene godina s minimalnom prosječnom pogreškom na skupu za testiranje od 4.8 godina. Zbog neuravnoteženosti skupa podataka i velike sličnosti ortopana osoba starijih od 60 godina dolazi do znatnijih pogrešaka u procjeni.

Svrha ovog rada, s obzirom da je identifikacija preminulih osoba često dug i težak posao koji obavljaju forenzički stomatolozi, je znatno olakšavanje i ubrzavanje tog procesa čineći ga ujedno znatno jeftinijim. Nadalje, smanjila bi se mogućnost ljudskih pogrešaka koje su izazvane visokom razinom emocionalnog stresa ovog posla. Dodatna prednost jest i jednostavna uporaba ovih metoda bez potrebnog predznanja ili posebne obuke.

Abstract (english)

The aim of this master thesis is to explore various algorithms for the analysis of information presented in dental X-ray images (orthopantograms) using deep learning. With an emphasis on the deployment of deep artificial neural networks inspired by the brain's biological neural networks, and the use of predetermined feature detectors.

An orthopantomogram or orthopan is an image of the complete upper and lower jaw with surrounding anatomical structures that provide general information on head bones, sinuses, jaws and teeth. Its use is to assess the degree of tooth maturation, order and speed of emergence and the dental age estimation.

Applicability is assessed on the dental age estimation from orthopantomograms as well as from partial human X-ray images. The results of the age estimation with the minimum mean error at the test set of 4.8 years have been achieved. Due to the imbalance of the data set and the large similarities of orthopedic images of persons over the age of 60, significant errors in the assessment result.

The purpose of this master thesis, since identification of deceased persons is often a long and difficult job done by forensic dentists, is to significantly facilitate and accelerate this process, making it much cheaper. Furthermore, the possibility of human error caused by the high level of emotional stress of this work would be reduced. An additional advantage is the simple use of these methods without the need of prior knowledge or special training.

• Keywords: duboko učenje; umjetna inteligencija; umjetne neuronske mreže; regresija; dentalna dob; stomatologija
• Keywords (english): deep learning; artificial antelligence; artificial neural networks; regression; dental age; dentistry
• Language: croatian
• URN:NBN: urn:nbn:hr:168:809792
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2020-01-07 17:42:01