The ubiquitous use of smartphones raises the need for stronger protection of these devices. Traditional authentication methods on mobile devices are still knowledge-based, exhibiting well-known drawbacks. In addition, requests for passwords, PINs, or screen lock patterns represent an interruption of the device usage. Unobtrusive biometric authentication techniques, e.g. background speaker or accelerometer-based gait verification, provide a means to improve the user experience. Since these methods only work within distinct environments, further unobtrusive authentication methods are required.
In this thesis the design and evaluation of a context-aware mobile biometric system is presented. An acquired database, which comprises contextual data of 26 different subjects, is employed for subject modeling and classification, and experiments are presented. Obtained results confirm the feasibility of the proposed system and provide hints for further optimizations.
Author: Heiko Witte