Project Abstract

This project is motivating by my interest in security of data in the modern world and investigating a novel idea of using EEG brain waves as a method of person recognition. This idea has only been around for a few years and research is continuing to be done on it, my research aims to optimise what is already established in terms of methods of preprocessing and creating an EEG Person recognition system, my project aimed to do this by reducing the inherent degradation present with EEG data between multiple sessions. The impact of this is to prove that through using a specific method outlined in my dissertation there are improvements on the accuracy of an EEG recognition system. The method I used was to first research other experiments done using similar data and follow what they did that worked whilst adding my methods of optimisation. I found that using Feature Alignment through Canonical Correlation did result in improvement between two machine learning models I trained for EEG person recognition thus proving that they reduced degradation between sessions. This means that this method could be added to future recognition systems using EEGs to make them more stable and accurate long term and over multiple sessions.

Keywords: Machine Learning, Biometrics, Cyber Security

 Conference Details

Session: Poster Session B at Poster Stand 54

Location: Sir Stanley Clarke Auditorium at Wednesday 8th 09:00 – 12:30

Markers: Scott Yang Yang, Monika Seisenberger

Course: BSc Computer Science FI, 3rd Year

Future Plans: I’m looking for work