An Electroencephalogram Based Detection of Hook and Span Hand Gestures

Brain-Computer Interfaces (BCI) make use of Electroencephalogram (EEG) signals to classify limb movements and other motor activities in various biomedical applications. This paper presents an EEG-based system to distinguish span and hook hand gestures. The proposed model consists of various signal processing techniques to extract features of interest and machine learning-based classification algorithms. We have extracted features based on statistical parameters calculated from the EEG readings. Fast Fourier Transform (FFT) along with the Windowing technique is implemented. 4 different classifying models namely Support Vector Machine (SVM), Adaboost, Decision Tree, and Random Forest, have been compared. The proposed method accurately classifies hook and span-hand gestures. The Random Forest classifier achieved the highest accuracy of 78.62% followed by Decision Tree and Adaboost.