| Brain computer interface (BCI) is one of the alternatives to assist severely disabled individual with no reliable muscular control. It is an interface which translates a human's neural activities into discrete computerized commands via electronic output devices. Recently, the neural activity acquired usually involves visual stimulation, motor imaginary, or physical motion. However, for those who have, or especially were born with physical impairment, the imagination of physical control can be difficult and unreliable. Using visual stimulation, therefore, is another alternative of interaction. Most BCI research, which is based on visual stimulation, used flickering LED light or images which the images were located at random positions and sizes on computer monitor, to elicit distinct neural activities. However, continuous uses this technique may cause eye fatigue, epileptic seizures, and visual impairment. The purpose of this study was to find an alternative which reduces these risks but remains as effective as previous techniques. We experimented with a classification type that uses color images on computer monitor that do not flicker. We compared the result of classification accuracy and completion time with the method of flickering images. In addition to the color classification, we experimented on a hybrid classification, which combined both techniques together. Furthermore, we would like to investigate the usability, such as user comfort and effectiveness of the classification type, when applying to a simple 2D menu. Color classifier resulted in significantly lower task completion time when selecting from menus. |
