Welcome to the InfoSci Platform
A Frequency Discrimination Technique for SSVEP-Based BCIs Using Common Feature Analysis and Support Vector Machine

A Frequency Discrimination Technique for SSVEP-Based BCIs Using Common Feature Analysis and Support Vector Machine

Akshat Verma, Praveen Kumar Shukla, Shrish Verma, Rahul Kumar Chaurasiya
Copyright: © 2022 |Pages: 21
ISBN13: 9781668439470|ISBN10: 1668439476|EISBN13: 9781668439487
DOI: 10.4018/978-1-6684-3947-0.ch009
Cite Chapter Cite Chapter

MLA

Verma, Akshat, et al. "A Frequency Discrimination Technique for SSVEP-Based BCIs Using Common Feature Analysis and Support Vector Machine." AI-Enabled Smart Healthcare Using Biomedical Signals, edited by Rahul Kumar Chaurasiya, et al., IGI Global, 2022, pp. 158-178. https://doi.org/10.4018/978-1-6684-3947-0.ch009

APA

Verma, A., Shukla, P. K., Verma, S., & Chaurasiya, R. K. (2022). A Frequency Discrimination Technique for SSVEP-Based BCIs Using Common Feature Analysis and Support Vector Machine. In R. Chaurasiya, D. Agrawal, & R. Pachori (Eds.), AI-Enabled Smart Healthcare Using Biomedical Signals (pp. 158-178). IGI Global. https://doi.org/10.4018/978-1-6684-3947-0.ch009

Chicago

Verma, Akshat, et al. "A Frequency Discrimination Technique for SSVEP-Based BCIs Using Common Feature Analysis and Support Vector Machine." In AI-Enabled Smart Healthcare Using Biomedical Signals, edited by Rahul Kumar Chaurasiya, Dheeraj Agrawal, and Ram Bilas Pachori, 158-178. Hershey, PA: IGI Global, 2022. https://doi.org/10.4018/978-1-6684-3947-0.ch009

Export Reference

Mendeley
Favorite

Abstract

BCI is a communication option that has come up as a very radical tool for those who are suffering from neuromuscular disorders. BCI provide a way for the brain to communicate with the outer world without the use of any outlying nerves. Steady state visually evoked potentials (SSVEP) are frequency-specific responses to visual stimuli. These are extensively used with EEG signals. This research projects an innovative method for recognition of SSVEP-based BCIs. The method establishes a processing pipeline where an IIR Butterworth filter is implemented which filters the signals that are further decomposed into waveforms also known as wavelets. Along with the wavelet decomposition, common feature analysis (CFA), canonical correlation analysis (CCA), and MCCA are used to extract features. The best result is obtained from DWT-CFA. The finest classification results are obtained from the RBF kernel-based SVM classifier. The best overall mean accuracy of 94.78% is obtained using DWT-CFA as the feature extraction technique and employing RBF kernel-based SVM as the classifier.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.