Epileptic Seizure Detection based on EEG signal analysis using Hierarchy based Hidden Markov Model

摘要

Epilepsy is defined as non communicable neurological disorder and characterized by repeated seizures. Electroencephalogram (EEG) is an efficient tool for analysing brain disorder. Low computation and efficient automatic epileptic seizure detection will be of great use for clinicians. In this research entropy features such as Shannon entropy, collision entropy and hjorth Factors are used as bio-marker for seizures detection. Feature selection is performed based on Analysis of variance (ANOVA) test. Efficiency of other features such as distance entropy and higuchi fractal dimension are evaluated. Hierarchy based Hidden Markov Model is used for classification. Two state ergodic Hidden Markov Model are designed for healthy-seizure, seizure-interseizure and healthy-interseizure classification. Average accuracy achieved for healthy-interseizure, healthy-seizure and seizure-interseizure are 95.62%, 96.67% and 95.00% respectively. The proposed algorithm is cross-validated with higher channel EEG signal.