Service Integration for Biometric Authentication
Name
Shazia Javed
Abstract
The success of biometric authentication systems is evident from the increasing rate of adoption of unimodal biometric systems in civil and governmental applications. However, this does not imply that biometric systems offer a complete authentication solution. Unimodal biometric systems exhibit a multitude of limitations which can be overcome by using multimodal biometric authentication systems. Multimodal systems are considered more reliable, and capable of meeting stringent performance needs and addressing the problem of non-universality and spoof attacks effectively. Despite the relative advantages, implementation and usability of multimodal biometric systems remain a fundamental software engineering challenge. Multimodal systems are usually an amalgamation of unimodal biometric systems chosen in accordance with the needs dictated by the business process(es) and the respective environment under consideration. The heterogeneity, availability of source code, and deployment needs for these systems incur significantly higher development and adaption costs. Being software engineers, we naturally strive to simplify the engineering process and minimize the required amount of effort. Therefore this work focuses on making the existing biometric systems reusable. The objective is to define a service integration framework which automates seamless configuration, and deployment of heterogeneous biometric systems, and minimizes the development effort and related costs. In this effort we replace the need for development and integration of scenario-specific compatible systems by repetitive scenario-specific configuration and deployment of multimodal biometric systems. The development of biometric systems is minimized to a one-time effort. We also present tools for configuration and deployment, which respectively configure and deploy multimodal biometric systems comprising of heterogeneous open source and/or commercial biometric systems required for fulfillment of domain specific authentication needs. In comparison to the prevalent practices, our approach reduces the effort required for developing and deploying reliable scenario-specific biometric authentication systems by 46.42%.
Graduation Thesis language
English
Graduation Thesis type
Master of Science in Engineering (4+2) Computer Science*
Supervisor(s)
Dr. Ulrich Norbisrath, Dr. Eero Vainikko
Defence year
2012