The application of neural network and spline wavelet models in the electroencephalogram analysis automation process

Abstract

The article focuses on the use of synthesized wavelets in the electroencephalogram analysis automation process. It describes the procedures for obtaining neural network and spline models proposed by the author. The advantages and disadvantages of the method are shown. The paper proposes a system of electroencephalogram analysis automation process based on the use of two levels of continuous wavelet transform. A detailed description of its operation is given. The paper describes a software package developed on the basis of the system. During the tests, the software feature detection accuracy (eye artifacts and pathological components) in the signal was 81.5%. It suggests the main areas for the developed system and package application, as well as ways for their further improvement.