PupilShutter: Annotation System Triggered by Pupil Constriction

Abstract

Various eye-controlled interfaces using voluntary action have been devel oped in recent years. In contrast to a major method using divergence eye movement as voluntary input, we focus on pupil constriction along with convergence. PupilShutter is an intelligent system that tells what you are gazing by detecting voluntary convergence eye movement with a wearable single eye-camera. When the pupil constriction is detected,a visual eld image is trimmed based on the gazing point and an annotation sound is produced by analyzing the trimmed image. In order to conrm the possibility of the use of pupil constriction as input, we measured the change of pupil diameter during near look operation. The result indicated that pupil constriction by near look operation can be used for the input of the system. Then, a user study was conducted to evaluate the usability of the system as an annotation interface by eye capture. As a result, the proposed method scored 99.2 % precision. Therefore, the system enables users to get information of the gazing object by near look operation when they want to know.