Epileptic seizure detection based on expected activity measurement and Neural Network classification.

Epilepsy is known as the second reason to visit a neurophysiologist after migraine. In this paper, we propose a new approach to automatically detect crises of epilepsy in an Electroencephalogram (EEG). Our algorithm is based on image transformation, Wavelet Decomposition (DWT) and taking advantage of the correlation between wavelet coefficients in each sub-band. Therefore, an Expected Activity Measurement (EAM) is calculated for each coefficient as a feature extraction method. These features are fed into back propagation Neural Network (ANN) and the periods with epileptic seizures and non-seizures are classified. Our approach is validated using a public dataset and the results are very promising, reaching accuracy up to 99.44% for detection epileptic seizures.