Pre-Processing Based Wavelets Neural Network for Removing Artifacts in EEG Data

Main Article Content

Dilipkumar Jang Bahadur Saini

Abstract

The electroencephalogram (EEG) is a record of brain activity; however, because the electric potential of cerebral activity has a low amplitude and occurs at frequencies between 4 to 60 Hz, it is easily masked by other environmental noise signals and non-cerebral signals. This work presents to minimize noise by pre-processing new wavelets which are numerically stable and orthogonal bases will be proposed using Morelette wavelets and classified using convolutional neural networks (CNN). For experimentation, wavelet transforms are done to the original EEG signals from the public EEG database using Python scripts. Performance measures like SNR and MSE, which are determined for various step sizes of signal and filter orders, are used to measure and analyse the performance of filters. Compared to existing methodologies, wavelet analysis techniques perform better.

Article Details

How to Cite
Jang Bahadur Saini, D. . (2022). Pre-Processing Based Wavelets Neural Network for Removing Artifacts in EEG Data. Research Journal of Computer Systems and Engineering, 3(1), 43–47. https://doi.org/10.52710/rjcse.40
Section
Articles