Facial Match : A step towards winning the battle against the fraudsters

Sammanfattning: Identity fraud is a severe and ruthless crime, regardless of leasing a car insomeone else’s name or illegitimately getting a loan from a bank. The protection depends strongly on technical development to further increase the safetyof ID checking.This project aims to design an extra security ID checking on top of humanobservation. With the help of an NFC Scanner which reads and scans ID cardsand passports, and the help of a web camera, an application is developed witha face recognition ability. The application displays a scanned face image andten newly captured face pictures of an individual. Thanks to robust face detection and recognition library called Luxand, the face of the individual willbe detected in the recently captured images and then compared with theID/passport face image.The project resulted in face recognition of an almost 90% success rate. Theresults were evaluated through 31 test cases and also by performing a hypothesis test. The statistical hypothesis test was performed to prove that there is asystematic benign effect of the proposed procedure with a p-value of0.000015, which indicates a high probability of success.