Cognitive and Expressive based Search Algorithm for Goggle Search Using BCI

  Ms. Nidhi Vala,  Prof. Kiran Trivedi

Generally Brain Computer Interfaces (BCI) require both the hardware and software support for making real communication alive. Emotiv Epoc Neuroheadset is one of those devices present in the market which provides this kind of communication through electroencephalograph (EEG) signals and is also used to explain the brain activities. By analysing the recorded EEG signals, characteristic patterns for various facial expressions, human emotions and cognitive actions can be identified. These patterns are very useful for the analysis of human behaviour in different situations & environment and for developing any real-time application. Here through analysis we have found the characteristic patterns for the expressive features eye blink and left smirk. Generally these BCI products are being used by the physically disable people as a tool replacing their immortal body element. But now non-invasive BCI products are gaining popularity among healthy people for their amusement and curiosity. We have developed an application which make use of the continuous raw EEG data captured from human brain through Emotiv Epoc Neuroheadset and the gained results of the two facial expressions for searching a particular string on Google. BCI virtual speller was one of the important application for severely disable people. But in this paper, it has been spotted that, the main challenge of designing a larger matrix based BCI speller can be solved in a different way. Here we have made use of the 5x10(RC) matrix where 41 symbols have been mapped in it. This matrix will use the expressive feature eye blinking as its left-to-right shift focus controller and left smirk as its top-to-bottom focus controller. As this technology is very new for the most of people we need to provide training before actual experiment. And that session was called cognitive training for the subjects. In our experiment we got the average accuracy up to 66.66% but in final session accuracy was 80%, plus typing speed was average 5 characters per 45.67 sec during the experiment. Performance got increased with sessions and it has been shown by graphical representation.

Keywords- Brain Computer Interface(BCI), Electroencephalograph(EEG), BCI Speller, RC Matrix - Row Column Matrix

Unique Identification Number - IJEDR1402242
Page Number(s) - 2805-2810
Pubished in - Volume 2 | Issue 2 | June 2014
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Ms. Nidhi Vala,  Prof. Kiran Trivedi,   "Cognitive and Expressive based Search Algorithm for Goggle Search Using BCI", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.2, Issue 2, pp.2805-2810, June 2014, Available at :http://www.ijedr.org/papers/IJEDR1402242.pdf