Home // ICCGI 2013, The Eighth International Multi-Conference on Computing in the Global Information Technology // View article
Authors:
Faraz Akram
Hee-Sok Han
Tae-Seong Kim
Keywords: P300; Brain Computer Interface; Word Typing; Human Computer Interaction
Abstract:
The conventional P300 brain computer interface (BCI) system for character spelling is typically composed of a paradigm that displays flashing rows or columns of characters and a P300 classifier from which a target character gets recognized. One significant drawback of this system in practice is its typing speed which could take a few minutes to type each character of a target word. In this work, we propose a novel BCI system through which a whole word can be typed with much higher word typing speed and accuracy. In our presented system, we have integrated a custom-built smart dictionary to give word suggestions upon few key characters initially typed by the user. Upon the suggested words, the user can select one out of the given suggestions to complete word typing. Our novel paradigm significantly reduces the word typing time and makes words typing more convenient. In the classification part, we have also adopted a new classifier, Random Forest (RF) instead of a commonly used Support Vector Machine (SVM). Our results with four subjects using the presented word typing system demonstrate an average typing time of 1.66 minutes per word, whereas the conventional took 2.9 minutes, improving the typing time by 42.75%. Also RF improves the P300 classification accuracy significantly, outperforming SVM. Our presented system could be useful for practical human computer interaction applications.
Pages: 106 to 109
Copyright: Copyright (c) IARIA, 2013
Publication date: July 21, 2013
Published in: conference
ISSN: 2308-4529
ISBN: 978-1-61208-283-7
Location: Nice, France
Dates: from July 21, 2013 to July 26, 2013