Target selection based on gaze likelihood in gaze input interface

In recent years, computer device operation by gaze input has become popular. Some of the advantages are that those devices can be operated without hand. People with disabilities, therefore, are able to operate computers. This is also suitable for elderly people for whom it is easier to use gaze input than mice. It is, however, difficult to point to a target steadily by using eye gaze as the input, because eyes have the property of involuntary movement. It is, therefore, sometimes hard to perform stable target selection and may cause target selection errors. In this paper, gaze point is regarded to spread spatially and temporally. This paper proposes an icon selection method based on gaze likelihood. The proposed method has realized to improve the target selection accuracy. As a result, target selections are succeeded extremely from 44.7% to 84.5% In conclusion, this paper shows a significance of the gaze likelihood.