Loading...

Epileptic Signal Denoising Using Morphological Component Analysis Based on Dictionary Learning

Ilmak Foroosh, Arman | 2021

168 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54827 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Shamsollahi, Mohammad Bagher
  7. Abstract:
  8. The prevalence of epilepsy in the world and the need for surgery to treat patients have made it essential to locate the site of epilepsy before surgery. One method is to apply source localization algorithms to the EEG signals of epileptic patients in the ictal and interictal periods. However, because these signals are contaminated with various noises, they are challenging to interpret and require noise cancellation. Therefore, various methods have been proposed to eliminate the noise. Among these methods, a new method recently used to remove noise from the epileptic signal is Morphological Component Analysis (MCA). This method uses the basic concepts of sparse representation of signals to eliminate noise. In this thesis, we intend to eliminate the noise of epileptic signals by using the pre-made dictionary and the trained dictionary, using the method of Morphological Component Analysis. Simulated data then evaluate the performance of these two methods. Furthermore, a source localization algorithm has been used to show the location of epileptic zone. Finally, the results of these methods on real data are also given. The final results show that using the method of Morphological Component Analysis with a trained dictionary can positively remove noise from the epileptic signal
  9. Keywords:
  10. Denoising ; Epileptic Seizure Detection ; Morphological Component Analysis ; Dictionary Learning ; Epilepsy ; Epileptic Brain Signal ; Sparse Representation

 Digital Object List

 Bookmark