title   
  

The first brain-computer interface utilizing a Turkish language model (Türkçe dil modeli kullanan ilk beyin-bilgisayar arayüzü)

Ulaş, Çağdaş and Çetin, Müjdat (2013) The first brain-computer interface utilizing a Turkish language model (Türkçe dil modeli kullanan ilk beyin-bilgisayar arayüzü). In: 21st Signal Processing and Communications Applications Conference (SIU 2013), Haspolat, Cyprus

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
326Kb

Official URL: http://dx.doi.org/10.1109/SIU.2013.6531174

Abstract

One of the widely studied electroencephalography (EEG) based Brain-Computer Interface (BCI) set ups involves having subjects type letters based on so-called P300 signals generated by their brains in response to unpredictable stimuli. Due to the low signal-to-noise ratio (SNR) of EEG signals, current BCI typing systems need several stimulus repetitions to obtain acceptable accuracy, resulting in low typing speed. However, in the context of typing letters within words in a particular language, neighboring letters would provide information about the current letter as well. Based on this observation, we propose an approach for incorporation of such information into a BCI-based speller through a Hidden Markov Model (HMM) trained by a Turkish language model. We describe smoothing and Viterbi algorithms for inference over such a model. Experiments on real EEG data collected in our laboratory demonstrate that incorporation of the language model in this manner leads to significant improvements in classification accuracy and bit rate.

Item Type:Papers in Conference Proceedings
Uncontrolled Keywords:P300 speller, Brain-Computer Interface, language model, Hidden Markov Model, Forward-Backward algorithm
Subjects:T Technology > TK Electrical engineering. Electronics Nuclear engineering
ID Code:23477
Deposited By:Müjdat Çetin
Deposited On:13 Jan 2014 15:29
Last Modified:28 Feb 2014 10:40

Repository Staff Only: item control page