抄録
A new input interface using eyelid movement associated with eye closure is proposed. The interface operates by combining reading information such as characters by synthesized voice and the selective switch using voluntary eyelid movement. This paper discusses the feasibility of hiragana input interface which a participant selects a designated hiragana during continuous reading hiragana by synthesized voice. Since eyelid movements contain involuntary eyelid movement, the interface must eliminate the response arising from involuntary eyelid movements to avoid false selection. By comparing the maximum amount (peak value) of voluntary eyelid movement with involuntary eyelid movement after each presentation of a hiragana, it is found that the former movement makes a significantly higher peak than the latter. This result indicates that discrimination of voluntary eyelid movement from involuntary eyelid movements based on the difference of peak value might be effective. Experiment on inputting a Japanese word constructed by four hiragana characters (e.g. “まつえし”) is conducted. Hit rate for each hiragana character involved in presented word exhibits 95% and false alarm rate for each hiragana character not involved in presented word shows 0.9%. High hit rate and low false alarm rate for hiragana characters confirm the possibility of the proposed method.
