Epileptic seizure Classification using EEG Signal

Abstract

Nowadays, there are many diseases which are very dangerous and cause death to our life since it may affect an important organs of a human body. One of such a dangerous disease is epileptic seizure. Since it will affect our brain and leads to death. So, there are many techniques have been proposed already to classify this disease. But those are not that much of an efficient to classify this epileptic seizure. Since, the mind is consistently dynamic with trade of electrical signals, which might be caught by electroencephalograph (EEG). In the EEG the electrical activity of neuron is reflected as like the wave pattern. These EEG signals have been contaminated by noise due to the environment factors. There is an extreme need to eliminate these noises prior to utilizing the wave as input to any diagnostic systems. So, this system proposed the utilization of daubechies wavelet (db8) to eliminate the noises and artefacts. In decomposition of eighth level there are Five frequency bands have been mined. The attributes such as minimum, variance, kurtosis, maximum, entropy, skewness, median, standard deviation, energy, frequency, mode, mean, phase magnitude have been extracted from the regenerated signal. The classifier will take these attributes as an input and classify the affected person from others.