doctoral thesis
Two-dimensional face image classification for distinguishing children from adults based on anthropometry

Petra Grd (2015)
Sveučilište u Zagrebu
Fakultet organizacije i informatike Varaždin
Metadata
TitleTwo-dimensional face image classification for distinguishing children from adults based on anthropometry
AuthorPetra Grd
Mentor(s)Miroslav Bača
Abstract
Classification of face images can be done in various ways. This research uses two-dimensional photographs of people's faces to detect children in images. Algorithm for classification of images into children and adults is developed and existing algorithms are analysed. This algorithm will also be used for age estimation. Through analysis of the state of the art researchon facial landmarks for age estimationand combination with changes that occur in human face morphology during growth and aging, facial landmarks needed for age classification and estimation of humans are identified. Algorithm is based on ratios of Euclidean distances between those landmarks. Based on these ratios, children can be detected and age can be estimated.
Parallel title (Croatian)Klasifikacija dvodeminezionalnih slika lica za razlikovanje djece od odraslih osoba na temelju antropometrije
Committee MembersMarkus Schatten (committee chairperson)
Miroslav Bača (committee member)
Kornelije Rabuzin (committee member)
Sandro Gerić (committee member)
Arun Ross (committee member)
GranterSveučilište u Zagrebu
Fakultet organizacije i informatike Varaždin
PlaceVaraždin
StateCroatia
Scientific field, discipline, subdisciplineSOCIAL SCIENCES
Information and Communication Sciences
Information Systems and Information Science
UDK004
GENERALLY
Computer science and technology. Computing. Data processing
Study programme typeuniversity
Study levelpostgraduate
Study programmePostgraduate doctoral study in Information Science
Academic title abbreviationdr.sc.
Genredoctoral thesis
Language English
Defense date2015-06-16
Parallel abstract (Croatian)
Slike lica mogu biti klasificirane na različite načine. Ovo istraživanje koristi dvodimenzionalne fotografije ljudskih lica za detekciju djece na slikama. Kreiran je novi algoritam za klasifikaciju fotografija ljudskih lica u dvije grupe, djeca i odrasli. Algoritam će se također koristiti za procjenu dobi osoba na slici te će biti analizirani postojeći algoritmi. Kroz analizu literature o karakterističnim točkama korištenih u procjeni dobi i kombinacijom dobivenih karakterističnih točaka s morfološkim promjenama tokom odrastanja i starenja, definirane su karakteristične točke potrebne za klasifikaciju i procjenu dobi. Algoritam se bazira na omjerima Euklidskih udaljenosti između identificiranih karakterističnih točaka.
Parallel keywords (Croatian)Biometrics face aging anthropometric ratios craniofacial morphology face changes Biometrija starenje lica antropometrijski omjeri kraniofacijalna morfologija promjene na licu
Versionaccepted version
Resource typetext
Access conditionOpen access
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
Noteaccepted version
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:211:517258
CommitterLjiljana Hajdin