prikaz prve stranice dokumenta Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika
Access restricted to authenticated users
undergraduate thesis
Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika
2022. urn:nbn:hr:179:604022
Tomić, Klara Bruna
University of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
Department of Electronics and Computing

Cite this document

Tomić, K. B. (2022). Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika (Undergraduate thesis). Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture. Retrieved from https://urn.nsk.hr/urn:nbn:hr:179:604022

Tomić, Klara Bruna. "Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika." Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2022. https://urn.nsk.hr/urn:nbn:hr:179:604022

Tomić, Klara Bruna. "Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika." Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2022. https://urn.nsk.hr/urn:nbn:hr:179:604022

Tomić, K. B. (2022). 'Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika', Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, accessed 28 November 2024, https://urn.nsk.hr/urn:nbn:hr:179:604022

Tomić KB. Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika [Undergraduate thesis]. Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture; 2022 [cited 2024 November 28] Available at: https://urn.nsk.hr/urn:nbn:hr:179:604022

K. B. Tomić, "Utjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika", Undergraduate thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, Split, 2022. Available at: https://urn.nsk.hr/urn:nbn:hr:179:604022

Please login to the repository to save this object to your list.
Metadata
TitleUtjecaj pretprocesnih koraka na performanse modela dubokog učenja u obradi medicinskih slika
Title (english)Influence of data preprocessing steps on the performance of deep learning models in medical image processing
AuthorKlara Bruna Tomić
MentorLinda Vicković https://orcid.org/0000-0002-9820-7960 (mentor)
GranterUniversity of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
(Department of Electronics and Computing)
Split
Defense date and country2022-07-08, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
U ovom je završnom radu metodologija umjetne inteligencije opisana kroz proces dubokog učenja. Predstavljene su umjetne neuronske mreže koje omogućavaju strojevima da u procesu dubokog učenja steknu inteligenciju nalik ljudskoj. Poseban je naglasak stavljen na konvolucijske neuronske mreže koje su svoju popularnost stekle zahvaljujući odličnim performansama koje pokazuju prilikom obrade slika. U sklopu završnog rada opisani su izazovi u radu s medicinskim slikama te je istraženo utječe li korištenje pretprocesnih koraka u njihovoj obradi na performanse modela dubokog učenja. Model dubokog učenja implementiran je koristeći Tensorflow Object Detection API. Za potrebe učenja modela prikupljene su MRI slike koljena na kojima je označen centralni dio koljena koji uključuje područje zone rasta na femuru i tibiji kojeg bi model trebao naučiti detektirati. Provelo se treniranje, evaluacija i testiranje različitih modela, a uspoređivanjem njihovih performansi stečen je uvid kako pretprocesni koraci i broj slika na kojima model uči utječu na detekciju objekta na medicinskim slikama koljena.
Abstract (english)
In this undergraduate thesis, the methodology of artificial intelligence is described through the process of deep learning. Artificial neural networks that enable machines to acquire human-like intelligence in the process of deep learning are presented. Special emphasis is placed on convolutional neural networks, which gained their popularity thanks to their excellent performance when processing images. As part of the thesis, the challenges in working with medical images were described and it was investigated whether the use of preprocessing steps in their processing affects the performance of deep learning models. The deep learning model is implemented using the Tensorflow Object Detection API. For the purposes of learning the model, knee MRI images were collected, on which the central area of the knee with the growth zones on the femur and tibia, which the model should learn to detect, was marked. Training, evaluation and testing of different models were carried out. By comparing their performances, insight into how the preprocessing steps and the number of images on which the model learns affect object detection on knee medical images was gained.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:179:604022
Study programmeTitle: 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 resourceText
File originBorn digital
Access conditionsAccess restricted to authenticated users
Terms of useLicence In copyright icon
Created on2023-09-28 12:04:24