Primjena algoritama strojnog učenja u prepoznavanju lica

Puljko, Benjamin

master's thesis

Primjena algoritama strojnog učenja u prepoznavanju lica

Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. urn:nbn:hr:200:060378

Puljko, Benjamin

Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Computer Engineering and Automation
Chair of Automation and Robotics

Cite this document

APA 6th Edition

Puljko, B. (2020). Primjena algoritama strojnog učenja u prepoznavanju lica (Master's thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:060378

MLA 8th Edition

Puljko, Benjamin. "Primjena algoritama strojnog učenja u prepoznavanju lica." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:060378

Chicago 17th Edition

Harvard

Puljko, B. (2020). 'Primjena algoritama strojnog učenja u prepoznavanju lica', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 16 June 2024, https://urn.nsk.hr/urn:nbn:hr:200:060378

Vancouver

Puljko B. Primjena algoritama strojnog učenja u prepoznavanju lica [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2020 [cited 2024 June 16] Available at: https://urn.nsk.hr/urn:nbn:hr:200:060378

IEEE

B. Puljko, "Primjena algoritama strojnog učenja u prepoznavanju lica", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:200:060378

Cite this item: https://urn.nsk.hr/urn:nbn:hr:200:060378

Please login to the repository to save this object to your list.

Metadata

Title	Primjena algoritama strojnog učenja u prepoznavanju lica
Title (english)	Application of machine learning algorithms in facial recognition
Author	Benjamin Puljko
Mentor	Emmanuel Karlo Nyarko (mentor)
Committee member	Emmanuel Karlo Nyarko (predsjednik povjerenstva)
Committee member	Ratko Grbić (član povjerenstva)
Committee member	Damir Filko (član povjerenstva)
Granter	Josip Juraj Strossmayer University of Osijek Faculty of Electrical Engineering, Computer Science and Information Technology Osijek (Department of Computer Engineering and Automation) (Chair of Automation and Robotics) Osijek
Defense date and country	2020-07-24, Croatia
Scientific / art field, discipline and subdiscipline	TECHNICAL SCIENCES Computing Artificial Intelligence
Abstract	U ovom radu ispitana su četiri algoritma strojnog učenja koja su se upotrebljavala za klasifikaciju. Prvotno je pohranjen skup podataka koji je zatim obrađen na način da su detektirana lica, izrezana, poravnata te pohranjena u podatak koji se kasnije učitava radi lakšeg izvršavanja eksperimenata. U prvom eksperimentu se za izdvajanje značajki i smanjenje dimenzionalnosti upotrebljavala analiza glavnih komponenata (PCA), dok je za drugi eksperiment upotrebljen facenet model za izdvajanje značajki. Zaključeno je da je facenet superiorniji u izdvajanju značajki od PCA. U trećem eksperimentu je također rabljen facenet model, ali s povećanim brojem klasa u odnosu na prvi i drugi eksperiment. U posljednjem, četvrtom eksperimentu, uspoređena je vgg16 naučena neuronska mreža s facenet naučenim modelom. Vgg16 model imao je točnost od 94,6%, dok je facenet model imao točnost od nevjerojatnih 99,8%. Zaključeno je da su neuronske mreže daleko učinkovitije u izdvajanju značajki te sam klasifikacijski algoritam ne igra veliku ulogu u točnoj klasifikaciji, ukoliko su značajke dobro izdvojene.
Abstract (english)	In this paper, four machine learning algorithms have been tested for classification. Initially, the dataset was processed in such a way that the detected faces were cut, aligned and stored in a file for later use. In the first experiment, principal component analysis was used for feature extraction and dimensionality reduction while for the second experiment, the pretrained facenet model was used. From the experiments, it was concluded that facenet is superior to PCA in feature extraction. In the third experiment, facenet model was used, but with a greater number of classes than in the first two experiments. In the last, fourth experiment, the vgg16 pretrained neural network was compared to the facenet pretrained model. Vgg16 model had an accuracy of 94,6% while the facenet model had an amazing accuracy of 99,8%. In conclusion, neural networks are far more efficient in feature extraction and the classification algorithm itself doesn't play a big role in correctly classifying the face, if the features are well extracted.
Keywords
Keywords (english)
Language	croatian
URN:NBN	urn:nbn:hr:200:060378
Study programme	Title: UNIVERSITY STUDY AUTOMOBILE COMPUTER AND COMMUNICATION Study programme type: university Study level: graduate Academic / professional title: magistar/magistra inženjer/inženjerka automobilskog računarstva i komunikacija (magistar/magistra inženjer/inženjerka automobilskog računarstva i komunikacija)
Type of resource	Text
File origin	Born digital
Access conditions	Open access
Terms of use
Repository	Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Created on	2020-09-15 15:15:52