EEG signal artifact removal using ORICA algorithm
Deepak Bansal, R. K. Sharma · 2017 International Conference on Trends in Electronics and Informatics (ICEI) · 2017
This work presents an implementation of online recursive independent component analysis (ORICA) processor for artifact removal of EEG signal. The system architecture consists of a covariance whitening unit, singular value decomposition (SVD) unit, an ORICA weight training unit and an auto de-artifact and reconstruct unit. The algorithm uses sample entropy feature for identifying eye blink artifacts after ICA decomposition. The design is implemented with 5-channel EEG signal dataset having sampling rate of 256 Hz. Each channel comprises of 1600 samples. Similarity between the artifactual EEG signal and clean EEG signal is obtained with average correlation coefficient of 0.9913.