Removal of Ocular Artifacts from EEG using Adjust Based ICA and Regression Algorithm

A portable brain imaging technique is electroencephalography (EEG)technology is used to capture brain activity. Blinking and eye movements cause significant artifacts with high amplitudes compared with normal EEG signal. It is crucial to removing these artifacts from EEG signals for better treatment. For removing artifacts, there are many algorithms used before. few of them are principal component analysis(PCA),wavelet transform. In this paper two techniques called adjust based ICA and regression algorithm are described to remove these artifacts from brain signals. For the first method, independent component analysis algorithm is used which contains major steps like mixing signals, data preprocessing, singular value decomposition and whitening. The signal containing mixture of EEG and EOG signals is taken and separated signals are shown in the results. Signal to noise ratio(SNR) for this method is 81dB,which is low. In order to give more SNR value, Regression algorithm is used. The SNR value observed in after executing this algorithm is 91.22dB which is more than adjust based ICA. In regression algorithm when the order is increased, the MSE is decreased.