English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The method of wavelet transformation and BP neural network for classification of epileptic intermission and epileptic attack signals was studied. The discrete binary wavelet transformation of EEG signal was carried out with db4 wavelet function. The third to fifth level signals by reconstruction and noise reduction were extracted as feature signals. The BP neural network with three levels was built for signal classification by the variation coefficient and fluctuation index. The accuracy recognition rate of EEG signals reaches 95% in simulation experiment. The results show that the method on classification of epileptic intermission and epileptic attack signals is effective and rapid. Introduction EEG signals was the voltage change of the cerebral cortex or the surface of the scalp which was caused by the brain electrophysiological activities[1],which was recorded through the electrodes. EEG signals containsed a large number of physiological and pathological information, it may provide the basis for the clinical diagnosis and treatment of brain disease through the analysis and processing of EEG signals.it is a non-stationary random signal under strong noise background,which contains the ingredients are very complicated. To diagnose the disease by using EEG signal,the effective information of singals must be extracted by using the appropriate signal processing methods for EEG classification]. At present there are many methods for EEG signal processing, including wavelet analysis method which is very suitable for noisy nonstationary EEG signalinformation extraction ,because it can make use of the multi resolution characteristics of multi-scale decomposition of signals,extracting time domain features band limited signal,at the same time to meet the requirements of analysis in frequency domain and time domain. According to the epileptic intermission and epileptic attack signals,this paper uses the method of wavelet transform to remove noise interference and extract Characteristics of EEG signal which are classified by BP neural network method,so as to identify the EEG signal. The wavelet transform of EEG signals Wavelet transform can decompose the EEG signal into multiple scales, Scales of effective signal and interference signal with different singularity can removal characteristics of noise and keep the effective signal. The main frequency range of epileptic EEG signals is between 3hz and 25hz, The high frequency signal is higher than that of 3-25hz reflected in the low scale after decomposition,can directly filter. Interference components ,whitch are overlapped with the signal frequency in other scales ,are filtered by setting the appropriate wavelet threshold. This paper uses the discrete binary wavelet transformation for signal processing, The treatment process is divided International Conference on Advances in Mechanical Engineering and Industrial Informatics (AMEII 2015) © 2015. The authors Published by Atlantis Press 285 into two steps: (1)The wavelet transform on EEG signals can get the layers of wavelet coefficients, according to the characters of brain electric signal to choose the appropriate wavelet functions and decomposition layers. (2) Select the appropriate threshold processing method to eliminate the noise in the low frequency wavelet coefficients of elements After denoising the scale wavelet coefficients of signals as the electrical characteristics, calculation process by Mallat algorithm quickly, in the process of main parameters for wavelet basis function and threshold, according to the following principles to select and Settings 1The wavelet basis function Used in signal processing of the wavelet basis function is usually consistent with the signal characteristics。Often used in signal processing of wavele function has dbN wavelet, symN wavelet and coifN wavelet,Use them for correlation calculation with epilepsy wave ,the results show that daubechies4(db4) wavelet maximum value associated with epilepsy wave, So, choose the db4 wavelet as eeg signals processing of wavelet base 2The threshold value Threshold selection methods are divided into hard threshold and soft threshold method. After soft threshold denoising process the signal is more smooth. There are four main threshold selection rules of Soft threshold method, sqtwolog,Stein, heursure, minimaxi), stein and minimaxi less risk, and are suitable for separation of weak signal from the noise, when the high frequency components of signals have overlap with noise. Beause the high frequency components of brain electrical signal exist overlap with noise, so choose threshold estimation with Stein can to avoid filtering effective signal. The eigenvalue extraction of EEG The amplitude of EEG signals of epileptic attack is usually higher than that of epileptic intermission .In addition, there exists obvious differences on the frequency of the signal fluctuation between epileptic attack and epileptic intermission. Therefore, use the variation coefficient and fluctuation index which can reflect the characteristics of the above signals as the characteristic value of classification of EEG signals, according to the following method to calculate respectively a formula to calculate of variation coefficient

Study on Classification of EEG Signals Based on Wavelet Transformation and BP Neural Network