Reworked Second Order Blind Identification and Support Vector Machine technique towards imagery movement identification from EEG signals
This work addresses noise removal in EEG signals for brain-computer interfaces, but it appears incremental as it combines existing methods with modifications.
The paper tackled the problem of identifying imagery motor movements from EEG signals by removing noise and classifying events, achieving improved classification accuracy with a reworked SOBI algorithm and SVM.
During imagery motor movements tasks, the so called mu and beta event related desynchronization (ERD) and synchronization (ERS) are taking place, allowing us to determine human patient imagery movement. However, initial recordings of electroencephalography (EEG) signals contain system and environmental noise as well as interference that must be ejected in order to separate the ERS/ERD events from the rest of the signal. This paper presents a new technique based on a reworked Second Order Blind Identification (SOBI) algorithm for noise removal while imagery movement classification is implemented using Support Vector Machine (SVM) technique.