Noise elimination in automatic detection of epileptic seizures by wavelet transform using feature selection algorithm

One of the most important symptoms of epilepsy is convulsions, whose detailed analysis is performed by electroencephalography (EEG) signal. Electroencephalogram, as a clinical tool to illustrate the electrical activities of the brain accurately, provides an appropriate method for diagnosing epilepsy disorders, which plays an important role in identifying this disease, especially seizures. Seizures resulting from epilepsy may have negative physical, psychological, and social consequences such as loss of consciousness and sudden death. With timely and correct identification of epilepsy, its effect can be treated with medicine or surgery. In this thesis, a brief review of the methods of identifying epilepsy using EEG signal analysis along with the separation of epileptic signals from healthy and normal signals has been done. Methods based on EEG analysis, from non-linear methods of signal processing, provide much better results due to the properties of signal dynamics