Classification of normal. focal, and generalized EEG signals using EMD and ANN

摘要

Epileptic seizure detection is performed by employing a clinical tool called Electroencephalography (EEG). Seizure detection by visual diagnosis is time consuming and a difficult task is therefore proposed to automatically classify normal, focal and generalized EEG signals in empirical mode decomposition (EMD). EMD breaks EEG in to various functions of intrinsic mode (IMF). Features like sample and fuzzy entropies are compressed by IMFs. The features are fed into the Artificial Neural Network (ANN) and the classification of fuzzy and sample entropies is compared in this study. It has shown that fuzzy entropy provides better discrimination of normal, focal and generalized. This method accomplished highest accuracy 99.44%, sensitivity 99.23%, and specificity 100% with fuzzy entropy.