Machine learning methods for EEG signal classification
Μέθοδοι μηχανικής μάθησης για την ταξινόμηση σημάτων ηλεκτροεγκεφαλογραφίας
View/
Keywords
EEG ; Neural network ; Classification ; Machine learningAbstract
Electroencephalography (EEG) together with a Brain-computer interface (BCI), is a common approach in brain monitoring and control of prosthetic devices. One common challenge in research applications is the cross-subject classification of the motor movement and motor imagery signals, as the EEG recordings are highly individualized. In this project, a Shallow 2D Convolutional Neural Network was developed, which was trained on lightly preprocessed EEG data. The data was provided by the PhysioNet EEG dataset. The dataset contained 109 subjects, with each subject having recordings of motor imagery and motor movements. The developed model in combination with the segmented data achieved high statistical results (2 classes: 80%, 4 classes: 58%), comparable to other results from research papers in this domain.
The code for this assignment is provided at the link bellow https://github.com/VasilhsXart/EEGClassification.git
Department
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
28Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού - Μη Εμπορική Χρήση - Παρόμοια Διανομή 3.0 Ελλάδα
Αναφορά Δημιουργού - Μη Εμπορική Χρήση - Παρόμοια Διανομή 3.0 Ελλάδα