Loading...

Seizure Detection in Generalized and Focal Seizure from EEG Signals

Mozafari, Mohsen | 2020

477 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 52972 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Hajipour, Sepideh
  7. Abstract:
  8. Epilepsy is one of the diseases that affects the quality of life of epileptic patients. Epileptic patients lose control during epileptic seizures and are more likely to face problems. Designing and creating a seizure detection system can reduce casualties from epileptic attacks. In this study, we present an automatic method that reduces the artifact from the raw signals, and then classifies the seizure and non-seizure epochs. At all stages, it is assumed that no information is available about the patient and this detection is made only based on the information of other patients. The data from this study were recorded in Temple Hospital and the recording conditions were not controlled, so there is a lot of artifact and noise on the data. On the other hand, this set of data includes a variety of focal and general epilepsies that make classification difficult. The proposed method for removing the artifact is based on blind source separation algorithms such as ICA, CCA and other methods such as EMD. A new algorithm has been proposed for classifying epochs based on clustering and classification. In the proposed algorithm, the recorded EEG channels are first divided into clusters at each epoch. Then the classification is done on the channels. In the next step, at each cluster, voting is done based on the predicted channels’ labels. If the score of a cluster is higher than a certain threshold, the epoch is selected as the candidate to receive the seizure label. In the final stage, a post-processing algorithm, which uses the spatial information obtained from seizure channels, is applied to the channel labels and the label of each time epoch is identified. In this study, the Leave One Subject Out method was used. By applying the mentioned algorithm 88.49% accuracy, 85.99% sensitivity, 90.12% specificity, and 85.03% precision were obtained
  9. Keywords:
  10. Epilepsy ; Seizure Detection ; Artifact Reduction ; Classification ; Electroencephalogram Signals Classification ; Electroencephalography

 Digital Object List

 Bookmark

No TOC