Farzaneh Shalbaf Hosseinabadi - School Of Electrical Engineering, Isfahan University Of Technology (IUT), Isfahan, Iran.
Amir HajiRassouliha - School Of Electrical Engineering, Iran University Of Science And Technology (IUST), Tehran, Iran.

The purpose of this paper is to investigate a novel algorithm to detect the start and stop time of the occurrence of seizure by observing one channel EEG signal. This algorithm has been designed in a way that it could simply be implemented in hardware in order to be used in real time applications. We have extracted some time domain as well as frequency domain characteristics in different ranges of frequencies with the aid of multi-stage FIR filter banks. The resultant feature matrixes are then passed through a pattern recognition artificial neural network to be classified in two categories of normal activity and epileptic activity. The accuracy of this algorithm tested for four cases is 98.92% for detecting the start time of the seizure (in second) and 99.64% for detecting the end time of the seizure (in second).

Seizure, EEG, Multi-Stage FIR Filter banks, Pattern Recognition Artificial Neural Network and Wearable Devices