| |
| Acadêmico(a): Matheus Luan Krueger |
| Título: Protótipo de Software para Autenticação Biométrica baseada na Estrutura da Íris em Dispositivos Móveis |
| |
| Abstract: |
| Biometric recognition is a widely used technique in access control of certain places or information. The use of human iris as a biometric key has shown robust and promising results. There are two control access ways: the first one, called verification, confirms wheter the biometric key belongs or not to the person whose name or user has been provided. Instead of asking for some name or user like it happens on the verification, the identification process uses only a biometric key to discover the owner. Some few steps are necessary to do the iris feature extraction and encoding, which are: acquisition, localization and normalization. This project presents the implementation of an application to perform the biometric authentication in Apple’s mobile devices platform, the iOS. The biometric sample’s acquisition is done using the device’s front camera and the built-in flash. In the iris circle localization step it was chosen to apply the Hough circle transform, while the feature extraction and encoding process made use of external classes. All the steps and methods concerning the application were evaluated using an internal iris database, resulting in an error rate of three point two per cent for intra-class comparisons and zero per cent for inter-class comparisons. Our iris database is composed only by light iris images due the limitation of the method implemented in the iris and pupil localization. |
| |
|
