Eye Blinking EOG Signals as Biometrics

• Published in: Biometric Security and Privacy pp. 121 - 140
• Main Authors: Abbas, Sherif N., Abo-Zahhad, M.
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2016; Springer International Publishing
• Series: Signal Processing for Security Technologies
• Subjects: Business mathematics & systems; Imaging systems & technology; Pattern recognition
• ISBN: 9783319473000; 331947300X
• ISSN: 2510-1498; 2510-1501
• DOI: 10.1007/978-3-319-47301-7_5

In this chapter, the feasibility of using eye blinking Electro-Oculo-Gram (EOG) signal as a new biometric trait for human identity recognition is tested. For this purpose, raw Electro-Encephalo-Gram (EEG) signals were recorded from 40 volunteers while performing the task of eye blinking. These signals were recorded using portable EEG headset, known as Neurosky Mindwave, which has wireless and dry electrodes at Fp1 position above the left eye. This makes it practical for biometric applications and for measuring EOG signals. For pre-processing, Discrete Wavelet Transform (DWT) is adopted to isolate EOG signals from brainwaves. Then, the onset and the offset of the eye blinking waveforms in the EOG signals are detected. After that, features are extracted using time delineation of the eye blinking waveform where important marks like the amplitude, position, and duration of the positive and negative pulses of the eye blinking waveform are employed as features. Finally, Discriminant Analysis (DA) classifier is used for classification. Moreover, a feature selection technique based on differential evolution is added for the proposed system. The best Correct Recognition Rate (CRR) achieved is 93.75 %. In verification mode, the lowest Equal Error Rate (EER) achieved is 7.45 %. Also, the permanence issue is evaluated using training and testing samples with different time separation between them. The optimistic results achieved in this chapter direct the scientific research to study different approaches for human identification using eye blinking to increase system’s performance. Moreover, eye blinking EOG biometric trait can be fused with other traits like EEG signals to build a multi-modal system to improve the performance of the EEG-based biometric authentication systems.