Vizijski sustav za identifikaciju osobe temeljen na analizi karakterističnih točaka lica

master's thesis

• Title (english): Visual system for person identification based on the analysis of characteristic points of the face
• Author: Karlo Posavec
• Mentor: Tomislav Stipančić (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2022-07-18, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Artificial Intelligence

Abstract

Za ljude najvažniji atribut za prepoznavanja nekog pojedinca je lice. Prepoznavanje lica pripada biometrijskim metodama poput prepoznavanja šarenice, prepoznavanja otiska prsta i prepoznavanja rasporeda vena. Biometrija postaje sve popularnija za korištenje kod identificiranja i provjere autentičnosti, zbog svoje visoke točnosti i jedinstvenosti za pojedinca, kao i praktičnosti stečene time što ne zahtijeva nikakav nametnuti teret kao što su tokeni ili zapamćene lozinke. Međutim, takvi načini identifikacije i provjere autentičnosti moraju biti zaštićeni od novih vektora napada. Kada se prepoznavanje lica koristi u svrhu provjere autentičnosti i identificiranja, jedan potencijalni vektor napada je lažno predstavljanje, odnosno predstavljanje nekog oblika lažne imitacije lica ovlaštenog korisnika. Uzimajući ovo u obzir unutar ovog rada razvijen je novi sustav evidencije pohađanja u realnom vremenu na temelju prepoznavanja lica s mogućnošću diferencijacije između pravog i lažnog lica. Razvijeno rješenje može služiti kao zamjena za tradicionalne sustave evidencije pohađanja u organizacijama i učionicama koje karakteriziraju dugo trajanje i nezgrapnost održavanja. Samo rješenje izrađeno je u programskom jeziku Python. Na kraju ovog rada provedena je evaluacija koja je uključivala ljudske subjekte.

Abstract (english)

For people, the most important attribute for recognizing an individual is the face. Face recognition belongs to biometric methods such as iris recognition, fingerprint recognition and vein pattern recognition. Biometrics are becoming increasingly popular for use in identification and authentication, due to their high accuracy and uniqueness for an individual, as well as the convenience gained by not requiring any imposed burden such as tokens or remembered passwords. However, such means of identification and authentication must be protected against new attack vectors. When facial recognition is used for authentication and identification, one potential attack vector is spoofing, that is, presenting some form of fake imitation of the face of an authorized user. Taking this into account, in this paper a new real-time attendance system based on face recognition with the ability to differentiate between real and fake faces has been developed. The developed solution can be used as a replacement for the traditional attendance systems in organizations and classrooms, which are characterized by a long duration and cumbersome maintenance. The solution was developed in the Python programming language. An evaluation involving human subjects was carried out at the end of this paper.

• Keywords: identifikacija osobe; prepoznavanje lica; detekcija lica; duboko učenje; računalni vid; sustav evidencije pohađanja; zaštita od lažiranja
• Keywords (english): person identification; face recognition; face detection; deep learning; computer vision; attendance system; anti-spoofing
• Language: croatian
• URN:NBN: urn:nbn:hr:235:034747
• Study programme: Title: Mechanical Engineering; specializations in: Design, Process and Energy Engineering, Production Engineering, Engineering Modelling and Computer Simulation, Marine Engineering, Industrial Engineering and Management, Materials Engineering, Mechatronics and Robotics, Autonomous Systems; Course: Mechatronics and Robotics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2022-07-13 08:15:26