H.K. GARG1, P.G. KOHLI2, A.K. KOHLI3
1ECE Department, Thapar University, Patiala-147 004, Punjab, India.
2Department of Physiology, Punjab Institute of Medical Sciences, Jalandhar- 144 001, Punjab, India.
3ECE Department, Thapar University, Patiala-147 004, Punjab, India.
Received : 31-12-2013 Accepted : 24-04-2014 Published : 28-04-2014
Volume : 2 Issue : 1 Pages : 41 - 48
World Res J Bioinformatics 2.1 (2014):41-48
Keywords : Electroencephalogram (EEG), Epileptic spike (ESs), Nonlinear energy operators (NEO), Nonstationarity, Teager energy operator (TEO)
Conflict of Interest : None declared
This correspondence compare the different evaluation techniques for detection of nonstationary epileptic spike (ES) in the electroencephalogram (EEG) signal. Time domain techniques are presented using the smoothed nonlinear energy operator (SNEO) based on the different time-domain window functions and adaptive threshold. The signal modeling approaches based on adaptive filtering like linear predictor, mid-predictor, end predictor, Kalman filter with and without ANN and smoothed Kalman fllter. The transform domain methods employed to reduce the noise by preprocess the signal using Singular value decomposition (SVD) and Wavelet transform (WT) techniques. The time and transform domain techniques (SNEO, wavelet and artificial neural networks) are collectively applied to get better performance.