Loading...

Classification of EEG Signals to Detect Predefined Words in Imagined Speech

Rajabli, Reza | 2016

662 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 49465 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Beigy, Hamid
  7. Abstract:
  8. Attention to the Brain Computer interfaces (BCI) because of their potentials in improving, enhancing and substituting daily task, especially in people who suffer from diseases, has been increasing in the recent years. Such systems, receive brain activities and by extracting suitable features, try to interpret the brain commands. The aim of this project is to explore the ability of electroencephalogram (EEG) signal for silent communication by means of decoding imagined speech in brain activities. The previous research results show that imagining a word in the mind causes changes in the brain signals. These changes are interchangeable among different words. As a result, discriminating between these changes can lead to detecting the word. To investigate this phenomena, in this research electroencephalogram signals are recorded from 12 different Persian-speaking subjects and attempts are made to detect the imagined words better using a proper feature selection method. Finally, the average accuracy obtained for different subjects is at least above 50% and at most below 80%. Also, for the further investigation of the proposed method, the method is evaluated on the available dataset of one of the papers. The results show that our method outperforms the method proposed by when the classifiers are the same because of the better performance of the feature selection algorithm that we have proposed
  9. Keywords:
  10. Brain-Computer Interface (BCI) ; Electroencephalography ; Brain Signal ; Words Detection ; Brain Signals Recording ; Dataset Gathering ; Imagined Speech

 Digital Object List

 Bookmark

...see more