A combined singular spectrum analysis and adaptive noise canceler method outperformed existing techniques in removing EOG artifacts from synthetic and real-life EEG signals.
A novel combined SSA and ANC technique effectively removes EOG artifacts from single-channel EEG signals, outperforming existing methods.
The electroencephalogram (EEG) signals represent the electrical activity of the brain. In applications, such as brain-computer interface (BCI), features of the EEG signals are used to control the devices. However, while recording, EEG signals often contaminated by electrooculogram (EOG) artifacts; such artifacts degrade the performance of the BCI. In this paper, we proposed a new technique using singular spectrum analysis (SSA) and adaptive noise canceler (ANC) to remove the EOG artifact from the contaminated EEG signal. In this technique, first, we proposed a novel grouping technique for SSA to construct the reference signal (EOG) for ANC. Later, using the extracted reference signal, the adaptive filter was employed to remove EOG artifact from the contaminated EEG signal. To quantify the performance of the proposed technique, we carried out simulations on synthetic and real-life EEG signals. In terms of relative root mean square error and mean absolute error, the proposed SSA-ANC method outperforms the existing techniques.
Maddirala et al. (Fri,) conducted a other in EEG signal contamination by EOG artifacts. Combined singular spectrum analysis (SSA) and adaptive noise canceler (ANC) vs. Existing techniques was evaluated on Relative root mean square error and mean absolute error. A combined singular spectrum analysis and adaptive noise canceler method outperformed existing techniques in removing EOG artifacts from synthetic and real-life EEG signals.