Detekcija uhljev s pomočjo konteksta

Oblak, Tim (author), Peer, Peter (mentor), Štruc, Vitomir (comentor)

Zaradi svojih biometričnih lastnosti uhlji predstavljajo zanesljiv in edinstven vir informacij, še posebej uporaben na področju identifikacije ljudi. Pogoj za uspešno prepoznavo uhlja je učinkovit način detekcije, ki kljub prekrivanju in različnim pozam objekte detektira z relativno velikim priklicem. V tej diplomski nalogi predstavimo nov način detekcije uhljev, ki z namenom potencialne izboljšave napovedovanja koristi informacijo o kontekstu obraza. Za potrditev domneve smo v začetku izboljšali eno od obstoječih metod detekcije. Rezultate slednje smo utežili z lokalizacijo potencialnih področij uhljev. Na koncu smo zgradili še lasten cevovod, ki informacijo o kontekstu prejme že na začetku. Rezultat končnega cevovoda je izrazito povečan priklic detekcij glede na posamezne slikovne točke, poleg tega pa ohranimo relativno dobro mero preciznosti. Na testni množici 250 slik tako dosežemo izboljšavo dodatnih 28,5 odstotnih točk po meri Jaccardovega koeficienta podobnosti.

• Language: Slovenian
• Keywords: biometrija, kontekst, uhelj, detekcija, globoko učenje
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2017
• PID: 20.500.12556/RUL-95923
• Publication date in RUL: 25.09.2017

OBLAK, Tim, 2017, Detekcija uhljev s pomočjo konteksta [online]. Bachelor’s thesis. [Accessed 5 June 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=95923

Secondary language

• Language: English
• Title: Context-aware ear detection

Abstract

Because of their biometric properties, ears provide a reliable and unique source of information that is useful in the field of human identification. The condition for successful ear recognition is an effective detection method, which, despite various occlusions and poses, detects objects with a relatively large recall rate. In this thesis we present a novel approach to ear detection, which uses additional face context information for potential prediction improvement. In order to confirm the presumption, we first improved one of the existing detection methods. The results of the latter are weighted using localization of potential ear areas. Finally, we designed our own pipeline, which uses face context information from the beginning. The result of the final pipeline is a significant increase in pixel-wise detection recall rate, while preserving a relatively high measure of precision. On our test set of 250 images, we achieve an improvement of 28.5 percentage points according to the Jaccard coefficient of similarity.

• Keywords: biometrics, context, ear, detection, deep learning