Blink input interface enabling multiple candidate selection through sound feedback

Abstract

Several input interfaces that employ eye-blinking detection have been proposed. These input interfaces can detect operating intentions indicated by eye-blink actions. In this study, we propose a new blink input interface with the aim of increasing the classifiable blink types in addition to the regular voluntary blink. The proposed interface employs a sound feedback that facilitates multiple candidate selection. The feedback sound represents each input candidate and is generated while switching over time. This interface enables the user to input their operating intentions by controlling the timing of eye opening. We developed a prototype with the proposed interface and evaluated the system. The results indicated that a total classification rate of 96.5% is achieved for four types of voluntary blinks with 10 subjects. This classification rate is in accordance that of our previous report for two types of voluntary blinks. Thus, we concluded that our proposed method is reliable.