Development of a Mobile EEG-Based Feature Extraction and Classification System for Biometric Authentication
Juris Kļonovs, Christoffer Kjeldgaard Petersen. Aalborg University Copenhagen
The aim of this work is to investigate the possibilities to build a mobile biometric authentication system based on electroencephalogram (EEG). The objectives of this work include the investigation and identification of the most feasible feature extraction techniques and how these features can be used for authentication purposes. Therefore, we review the relevant literature, conduct several EEG measurement experiments and discuss their procedure and results with experts in the EEG and digital signal processing (DSP) fields. After gaining enough knowledge on the feature extraction and classification techniques and proposing the most applicable ones for our problem, we build and present a mobile prototype system capable of authenticating users based on the uniqueness of their brain-waves. Furthermore, we implement a novel authentication process, which leads the authentication system to be more secure. We also assess the usability of the system and define possible usage scenarios and propose a number of practical suggestions for future improvements of the system.