Title: Research on dry electrode SSVEP classification algorithm based on improved convolutional neural network

Authors: Mengjia He; Yingnian Wu; Tao Wang; Yujie Chen

Addresses: College of Automation, Beijing Information Science and Technology University, Beijing, China ' College of Automation, Beijing Information Science and Technology University, Beijing, China ' College of Automation, Beijing Information Science and Technology University, Beijing, China ' College of Automation, Beijing Information Science and Technology University, Beijing, China

Abstract: SSVEP signal is mainly collected by a wet electrode. Wet electrode collection steps are tedious, and dry electrode collection is simple and easy to apply, so the study of dry electrode is the focus of SSVEP research. Although the dry electrode is easy to collect, the ITR is low and the signal intensity of the EEG signal of the subjects is highly differentiated. At present, the cross trial classification of SSVEP is the focus and difficulty of research. In order to be able to meet the situation without calibration and can still have higher adaptability, in this paper, deep convolutional neural network is improved for the classification of dry electrode SSVEP. The classification of ten types of SSVEP dry electrode signals can reach the classification accuracy of 91.11%, and the ITR of 1S can reach 54.87, which has a very broad application scenario.

Keywords: BCI; steady-state visual evoked potential; SSVEP; dry electrode; convolutional neural network; no calibration.

DOI: 10.1504/IJSCOM.2021.114664

International Journal of Service and Computing Oriented Manufacturing, 2021 Vol.4 No.1, pp.70 - 88

Received: 21 May 2020
Accepted: 18 Jul 2020
Published online: 29 Apr 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article