Automatic detection and classification of EEG artifacts using fuzzy kernel SVM and wavelet ICA (WICA)
- PDF / 1,143,655 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 11 Downloads / 254 Views
(0123456789().,-volV)(0123456789().,-volV)
METHODOLOGIES AND APPLICATION
Automatic detection and classification of EEG artifacts using fuzzy kernel SVM and wavelet ICA (WICA) K. Yasoda1 • R. S. Ponmagal2 • K. S. Bhuvaneshwari3 • K. Venkatachalam4
Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Electroencephalography (EEG) is almost contaminated with many artifacts while recording the brain signal activity. Clinical diagnostic and brain computer interface applications frequently require the automated removal of artifacts. In digital signal processing and visual assessment, EEG artifact removal is considered to be the key analysis technique. Nowadays, a standard method of dimensionality reduction technique like independent component analysis (ICA) and wavelet transform combination can be explored for removing the EEG signal artifacts. Manual artifact removal is timeconsuming; in order to avoid this, a novel method of wavelet ICA (WICA) using fuzzy kernel support vector machine (FKSVM) is proposed for removing and classifying the EEG artifacts automatically. Proposed method presents an efficient and robust system to adopt the robotic classification and artifact computation from EEG signal without explicitly providing the cutoff value. Furthermore, the target artifacts are removed successfully in combination with WICA and FKSVM. Additionally, proposes the various descriptive statistical features such as mean, standard deviation, variance, kurtosis and range provides the model creation technique in which the training and testing the data of FKSVM is used to classify the EEG signal artifacts. The future work to implement various machine learning algorithm to improve performance of the system. Keywords Wavelet ICA (WICA) Fuzzy kernel support vector machine (FKSVM) Aircraft ECG signal
1 Introduction Communicated by V. Loia. & K. Venkatachalam [email protected] K. Yasoda [email protected] R. S. Ponmagal [email protected] K. S. Bhuvaneshwari [email protected] 1
Department of Biomedical Engineering, SNS College of Technology, Coimbatore, India
2
School of Computing, SRM Institute of Science and Technology, Kattankulathur, Kanchipuram, Tamilnadu, India
3
Department of Computer Science and Engineering, Karpagam College of Engineering, Coimbatore, Tamilnadu, India
4
School of CSE, Vellore Institute of Technology University Bhopal, Bhopal, India
The brain signal is recorded with electroencephalography (EEG) method in which electrical activity of the cerebral cortex is monitored and different electrodes are placed on the scalp. Presently, noninvasively, an electroencephalography signals are recorded and monitored. Clinical diagnosis and sleep disorders are most widely identified by the EEG technique. Data preprocessing is required when the visual inspection artifact is not a final one, and these artifacts may go ahead with ambiguous results. Generally, the segmentation of the whole affected part with the artifacts is difficult to classify that may in turn leads irrelevant data
Data Loading...
