Raspoznavanje lica iz video zapisa na osnovu jednog uzorka za treniranje po osobi

scientific master's theses

• Author: Marijan Živić
• Mentor: Mario Kovač (mentor)
• Committee member: Mario Kovač (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Control and Computer Engineering); Zagreb
• Defense date and country: 2011, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Ovaj magistarski rad opisuje sustav za raspoznavanje lica iz video zapisa osobe. Osnovni uvjet sustava je korištenje jedne frontalne slike lica osobe u postupku treniranja. Frontalne slike lice su snimljene profesionalnim fotoaparatom i osvjetljenjem, slične su fotografijama koje se koriste za izradu identifikacijskih dokumenata. Sustav rješava problem poznat po nazivu „problem jednog uzorka za po osobi“. Predložen je novi sustav za rješavanje problema jednog uzorka po osobi. Sustav pronalazi preciznu lokaciju zjenica pomoću prstenastih Gaborovih filtera na osnovu koje se slike lica geometrijski normaliziraju. Na geometrijski normaliziranim slikama lica se izvodi normalizacija osvjetljenja. Prikazane su usporedbe metoda normalizacije osvjetljenja pomoću lokalnih binarnih uzoraka i gradijentnih lica. U provedenim eksperimentima je prikazana veća pogodnost metode gradijentnih lica. Također je pokazano kako sustavi raspoznavanja temeljeni na KPCA metodi daju značajno bolje rezultate na realnim video zapisima.

Abstract (english)

This master thesis represents face recognition system in video. The system is based on one image with frontal pose in training procedure. Training images with frontal faces were taken with professional camera and lightning. Training images are similar to those used in identification documents. System solved problem know as one sample per person problem. A new recognition system is represented. System first does localization of eye pupil center. Using centers of both eye pupils system does the geometric normalization. Local light normalization is then applied on geometrically normalized images. Comparison of different light normalization techniques is represented. The performed experiments show better behavior of gradient faces as light normalization technique. It has also been demonstrated that KPCA recognition method is more suitable in recognition performed on video clips that simulate real situations.

• Keywords: biometrija; video; lokalizacija; lice; PCA; KPCA
• Keywords (english): biometrics; video; localization; face; PCA; KPCA
• Language: croatian
• URN:NBN: urn:nbn:hr:168:589348
• Study programme: Title: Postgraduate master programe in computing; Study programme type: university; Study level: postgraduate; Academic / professional title: magistar inženjer računarstva (magistar inženjer računarstva)
• Catalog URL: http://lib.fer.hr/cgi-bin/koha/opac-detail.pl?biblionumber=36997
• Type of resource: Text
• Extent: 95 str ; 30 cm
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2020-08-26 08:11:19