EEG analysis using steady-state visually evoked potentials for the real-time intention recognition

Abstract

In this research, the integrated user interface system has been developed to recognize the subject’s intention for telecommuting to the other device in real-time. The system is composed of three parts; the algorithm set of real-time EEG signals using steady-state visually evoked potentials(SSVEP) from visual stimulus, the communication system between the EEG data and algorithm set, and the combined vision system of web-cam based 3D space and visual stimulus set. The algorithm set is developed based on the canonical correlation analysis operated by using the well-known Python library, scikit-learn, and gives a result of gazing position of the subject calculated from just before four seconds of EEG dataset measured in the occipital lobe. The communication system is simply made by Lab streaming layer(LSL) which allows to transmission of EEG data from the EEG equipment to the algorithm set and receiving the obtained gazing position of the subject. In addition, the combined vision system is composed of real-time 3d space read by webcam and the visual stimulus experimental setup.