A novel peak signal feature segmentation process for epileptic seizure detection
- PDF / 1,001,775 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 22 Downloads / 184 Views
ORIGINAL RESEARCH
A novel peak signal feature segmentation process for epileptic seizure detection T. Perumal Rani1
•
G. Heren Chellam1
Received: 1 November 2019 / Accepted: 26 September 2020 Ó Bharati Vidyapeeth’s Institute of Computer Applications and Management 2020
Abstract Epilepsy is a brain disease in nerves which causes sudden seizure, sensations, and once in a while loss of mindfulness. This disorder is difficult to find manually because of its unpredictable nature since it is very hard to treat. The World Health Organization states that fifty million people having this type of disorder worldwide. Automatic detection assumes a significant role in the finding of epilepsy for it can get imperceptible data of Epileptic Electroencephalogram Signals precisely and diminish the burdens of medical field. The Brain’s function is monitored by using these EEG signals electrically. The goal of this paper is to find a classification on Electroencephalogram (EEG) signals using the Bonn University datasets. In order to address this challenge, we propose a new Peak Signal Features (PSF) method which extracts high and low peak features from EEG signals. In addition, Support Vector Machine, Decision Tree and K-Nearest Neighbor are used for classification. Finally, overall accuracy and the Mean Square Error rates of the above three classification methods with proposed method are measured. The experimental result demonstrates the effectiveness of the proposed approach. It also proves that SVM with proposed Peak Signal Features method gives better result than the other methods.
& T. Perumal Rani [email protected] G. Heren Chellam [email protected] 1
Department of Computer Science, Rani Anna Government College for Women (Affi. Manonmaniam Sundaranar University), Abishekapatti, Tirunelveli 627008 Tamilnadu, India
Keywords Butter worth filter Classification Decision tree Electroencephalogram Entropy Epilepsy Knearest neighbor Machine learning Support vector machine Abbreviations PSF Peak signal features SVM Support vector machine DT Decision tree KNN K-nearest neighbor ESD Epileptic seizure detection
1 Introduction Epilepsy is one of the challenging diseases which causes recurrent seizures and reoccur repeatedly. Imbalanced electrical change of the EEG signal may indicate abnormal by the electrodes which are placed on the scalp to record the brain activity. Abnormalities in these signals represent brain diseases like dementia, mental disorder, epilepsy, stroke, etc. [1]. By using these signals, epilepsy can be found out easily. Finding this type of disease on EEG signals manually is a tedious one and it takes time. Automated systems are used to detect normal and abnormal patterns easily. The proposed Peak Signal Features (PSF) extracts the high and low peak signal features. Butter worth filter is used to remove noise from the input EEG signal. Segmentation performance selects high peak and low peak signal both horizontally and vertically. Adaptive threshold method is used for high and low peak sign
Data Loading...
