EEG Channel Selection for Person Identification Using Binary Grey Wolf Optimizer

Electroencephalogram signals (EEG) have provided biometric identification systems with great capabilities. Several studies shown that EEG introduces unique and universal features besides specific strength against spoofing attacks. Essentially, is a graphic recording of the brain’s electrical activity calculated by sensors (electrodes) on scalp at different spots, but their best locations are uncertain. In this paper, channel selection problem formulated as binary optimization problem, where version Grey Wolf Optimizer (BGWO) used to find an optimal solution for such NP-hard problem. Further, Support Vector Machine classifier Radial Basis Function kernel (SVM-RBF) then considered EEG-based person identification. For feature extraction purposes, we examine three auto-regressive coefficients. A standard motor imagery dataset employed evaluate proposed method, including four criteria: (i) Accuracy, (ii) F-Score, (iii) Recall, (v) Specificity. experimental results, method (named BGWO-SVM) obtained 94.13% accuracy using only 23 5 Besides, BGWO-SVM finds electrodes not too close each other capture relevant information all over head. As concluding remarks, achieved results concerning number selected channels competitive classification accuracies meta-heuristics algorithms.