Epileptic seizure detection from EEG signal using Flexible Analytical Wavelet Transform

Epileptic seizure is the abnormal synchronous neuronal activity that occurs in human brain. The early detection of epileptic seizure helps in improving patient's mental health. In this work, an Electroencephalogram based methodology of automated epileptic seizure detection using Flexible Analytical Wavelet Transform is presented. In the proposed methodology, Electroencephalogram signals are decomposed into approximate and detailed wavelet coefficients using Flexible Analytical Wavelet Transform, initially. The statistical features such as mean, kurtosis and skewness are calculated from the selected wavelet coefficients as features. Further, the features are fed to the soft computing techniques for classifying Electroencephalogram data in seizure and non-seizure classes. Three soft computing techniques such as Support Vector Machine, Artificial Neural Network and Random Forest Tree classifiers are used for classification. The results of the classification yield the efficacy of proposed methodology of feature extraction in automatic epileptic seizure detection.