MSBE Analysis with Power Metric for Automated Identification of Epileptic Seizure
Hemlata Pal1, Abhay Kumar2
1Hemlata Pal*, School of Electronics, Devi Ahilya University, Indore, India.
2Dr. Abhay kumar*, Head of Department of School of Electronics, Devi Ahilya University, Indore, India.
Manuscript received on September 12, 2019. | Revised Manuscript received on October 20, 2019. | Manuscript published on October 30, 2019. | PP: 1-5 | Volume-9 Issue-1, October 2019 | Retrieval Number: A1000109119/2019©BEIESP | DOI: 10.35940/ijeat.A1000.109119
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Objective-This study explores a novel application of multi scale bubble entropy analysis with power metric analysis to achieve efficient epileptic seizure prediction performance. Method-This paper aims to develop a reliable seizure detection technique that incorporates AM FM model for decomposition of EEG into different sub bands. The initially first feature set is formed by acquiring the absolute and relative power components at each electrode. Second feature set is constructed by multi scale bubble entropy analysis from each sub band. These two major feature vectors are fuse into an integrated feature space to perform classification task using ANN. Result-Experimental results show that this method presents: 1) Consistent increase in complexity measures, 2) Increase in stability & discrimination of power. These finding suggest that extracted features can be used for treatment of epilepsy. Significance- This method provides greater stability, so this technique could be used to detect wider range of seizures.
Keywords: Epilepsy, EEG, Multi scale bubble entropy, Relative power, Seizure detection.