Automatic Detection of Epileptic Seizures Based on Entropies and Extreme Learning Machine
Epilepsy is a common neurological disease, and it is usually judged based on EEG signals. Automatic detection and classification of epileptic EEG gradually get more and more attention. In this work, we adopt two-step program to implement the automatic classification. Three entropies (approximate entropy, sample entropy, permutation entropy) are extracted as features to prepare for classifying. Then extreme learning machine is utilized to realize feature classification. Experimental results on Bonn epilepsy EEG dataset indicate that the proposed method is capable of recognizing normal, pre-ictal and ictal EEG with an accuracy of 99.31%, which is helpful for doctors to diagnose epilepsy disease.
KeywordsEntropy Extreme learning machine EEG signal classification
This work was supported by the project (61374154) of the National Nature Science Foundation of China and the Fundamental Research Funds for the Central Universities (DUT16RC (3)123).
- 4.Han, M., Sun, L.: EEG signal classification for epilepsy diagnosis based on AR model and RVM. In: IEEE ICICIP, pp. 134–139 (2010)Google Scholar