An Experimental Investigations on Classifiers for Brain Computer Interface (BCI) based Authentication

摘要

Brain signal based authentication system is relatively new when compared to other types of biometric data. This paper is based on Brain-Computer Interface (BCI) technique using Electroencephalography (EEG) signals to authenticate a system for paralyzed persons. Brain-computer interface (BCI) is a direct communication between computer(s) and the human brain. It is a system that facilitates the external device control by using the signals measured from the brain. BCI measures brain activity and by the way, those different brain signals are translated into control signals. An electroencephalograph (EEG) would read those brain waves using non-invasive electrodes that record the signals. The unique patterns can be used as a password or biometric identification. When we perform mental tasks like picturing a shape or performing an event, our brain generates unique neuronal electrical signals. Individuals are asked to perform some events for specified duration and their signals are acquired. These acquired signals are pre-processed, and the necessary time domain and frequency domain features are extracted and given to the classifiers for training. These classified signals are used for authentication purpose.