Laborpraktikum: Multimodal Biometrics
Dozenten:
Dr. Claus Vielhauer
Andreas Lang
Frank Zöbisch
Vorlesungszeit:
Di. 17:00 - 19:00, G29-146
Der Kurs startet am 17.04.2007
Anmeldung:
Bitte registrieren Sie sich hier.
Beschreibung:
Multimodal biometric systems use combinations of various biometric modalities, algorithms, sensors and/or instances in order to compensate for problems in single-modality biometric approaches (e.g. noisy data, intra-class variations, inter-class similarities, non-universality, spoof attacks). Further higher authentication performance, security levels and user acceptance can be achieved using such fusion scenarios or a combination of them. Despite to additional handling complexity (e.g. presentation of more than one biometric characteristic to more than one sensor) is that problems caused by disabilities, illness and/or age can be balanced by multimodal biometric approaches (particularly scenarios 3 and 4). For example, a missing trait or a trait that can be recognized poorly can be ignored.
Tasks of this course comprise the following areas:
- Implementation of normalization, fusion and weighting strategies
- Implementation of tools (acquisition and/or authentication) for additional modalities (e.g. finger print, hand print, …)
- Acquisition of biometric data