Automatic epileptic seizure detection based on the discrete wavelet transform approach using an artificial neural network classifier on the scalp electroencephalogram signal

Epilepsy is a common neural abnormality of the nervous system, which is constituted by recurring abnormalities, and humans with their class often suffered from disgrace and inequity. Therefore seizure detection has much importance in the medical domain for the treatment of epilepsy-affected persons. Electroencephalography signals are the most commonly used data modality for epilepsy identification. Yet, it can be challenging to proclaim the understated but severe modulation present in electroencephalogram signals. Here discrete wavelet transform is used for feature extraction, and the extracted features are classified by using an artificial neural network machine learning classifier. The 12 features are classified, and the accuracy among them is compared. To estimate the efficiency of the proposed method, different classifiers and feature extracting techniques are considered in this work. The obtained outcomes have proved that the new system obtained a promising identification accuracy of 97.24%, that is, capable of being an important approach for realistic applications curing epileptics.