Research Article
Personalized EEG Feature Extraction Method Based on Filter Bank and Elastic Network
@INPROCEEDINGS{10.1007/978-3-030-57115-3_10, author={Jian-Guo Wang and Zeng Chen and Yuan Yao}, title={Personalized EEG Feature Extraction Method Based on Filter Bank and Elastic Network}, proceedings={Bio-inspired Information and Communication Technologies. 12th EAI International Conference, BICT 2020, Shanghai, China, July 7-8, 2020, Proceedings}, proceedings_a={BICT}, year={2020}, month={8}, keywords={Brain Computer Interface (BCI) Motor imagery Elastic net Feature extraction}, doi={10.1007/978-3-030-57115-3_10} }- Jian-Guo Wang
Zeng Chen
Yuan Yao
Year: 2020
Personalized EEG Feature Extraction Method Based on Filter Bank and Elastic Network
BICT
Springer
DOI: 10.1007/978-3-030-57115-3_10
Abstract
In the practical application of the Brain Computer Interface (BCI) system, because of the diversity between the individuals in the electroencephalogram (EEG) system, the manifestation of Brain signals of each individual is different, so it is necessary to conduct personalized screening for different individuals to obtain information that is conducive to the classification of the EEG signals of the movement imagination. Because the EEG signal manifestation and corresponding rhythm range of different individuals are different, and the EEG characteristics corresponding to different frequency bands are also different, this paper proposes a personalized feature extraction method based on filter bank and elastic network. Based on several commonly used feature extraction and classification algorithms in the current BCI system, the analysis and research are carried out. The best combination method to obtain higher calculation rate and recognition accuracy provides some theoretical reference for the practical application of BCI system. Thus, the shortcomings of the CSP algorithm with better feature extraction effect are improved, and the proposed method can eliminate the individual differences of EEG signals, realize automatic feature selection, and improve classification accuracy.
