Balneo and PRM Research Journal, Journal Year: 2024, Volume and Issue: 15(Vol.15, no. 4), P. 763 - 763
Published: Dec. 22, 2024
Motor imagery electroencephalogram based brain computer interface systems can help people with disabilities to communicate an external device and realize rehabilitation therapies. The paper proposes flexible analytic wavelet transform (FAWT) as feature extraction method. method was tested on a dataset that contains EEG signals acquired from subjects motor disabilities. Classifiers linear discriminant analysis (LDA), quadratic (QDA), k nearest neighbors(kNN), Mahalanobis distance (MD) support vector machine (SVM) were utilized classsify the extracted features of right hand feet (FEET). best performance given by QDA classifier classification rate 97 %, sensitivity 99.65%, specificity 98.47%, kappa coefficient 0.97 F1 score 0.98. proposed shows through obtained results be used easy implement for assisting rehabitation real time BCI systems.
Language: Английский