Willer Travassos

A Practical Face Recognition System Using a Game with a Purpose: Basics, Concepts, Methods Willer Travassos

A facial recognition system is a computer application built to automatically identify or verify the identity of a person from a digital source. The quality of the source and environment from which digital information is retrieved pose problems to a face recognition system (FRS) that lead to erroneous results. Even though there is a necessity for systems that are capable of performing facial recognition on the fly, current systems, in order to be accurate, try to completely or partially control the environment of its digital sources. In this thesis, we approach the problems that afflict these systems by using a "game with a purpose" (GWAP). In our GWAP-based approach, we create an online game that uses human vision to perform facial recognition. We use the aid of humans because we still hold the edge over FRSs at recognizing faces and their features. It is through a GWAP that we channel our superior visual skills to deal with problems that affect current FRSs.