Adaptive Symmetrical Virtual Keyboard Based on EOG Signal

Abstract

Electrooculography (EOG) signal is a bio-signal that could play important roles in building communication between human and machine. This bio-signal is extracted from gaze motions by attaching electrodes to the skin around the eye. This paper presented the design and development of gaze-motion controlled adaptive virtual keyboard. The keyboard has the ability to arrange character layout based on specified frequency of a character. The designed virtual keyboard has 45 keys that comprising 26 letters, 10 numerals, 4 symbols, a space, a delete, an enter, a return and a central point key. The keyboard layout has six zones. Characters in the same zone have equal distance to the central point. The distance is the least gazing that possible to move from the central point to the character. The most frequently used character will shift to the shortest distance position. Seven paragraphs were typed using this virtual keyboard. The keywords of each paragraph are ‘bukittinggi’, ‘zimbawe’, ‘andalas’, ‘sukarno’, ‘randang’, ‘kapal’ and ‘dokter’. The performance of the adaptive virtual keyboard was evaluated using these seven keywords. Results shows that the number of gaze motion required to type the keyword after the re-arrangement of layout decreased about 27%. This adaptive ability can reduce the number of gazing to prevent eye fatigue.