研究実績の概要 |
The objective of the project is to identify the best dwell time for individual users of authentication applications using eye tracking. The goal is to obtain data from a large variety of users, including elderly persons, regarding the “Midas Touch problem”: how can we avoid object selection mistakes during eye tracking that occur due to an inappropriate dwell time? Details and the first results of the research are described in Paulus and Remijn (2021). Usability of various dwell times for eye-gaze-based object selection with eye tracking, (2021). Displays, 67, 101997. In the first half of the research stage, users (students) were asked to use their eye gaze to select 3 types of objects from a display. The objects consisted of alphanumeric characters, dot patterns, and visual icons. Object selection time, success rate, selection corrections, and dwell time evaluations were obtained. The main result was that for each object type, an object dwell time of 600 ms was preferred. The next step in the project is to investigate the generalizability of the results. For this purpose, we are performing a large-scale experiment using the exact same method as reported above. Data are obtained from three different user populations. So far data have been collected from 50 participants: 22 younger participants (16 to 35 years old), 20 mid-aged participants (36 to 55 years old), and 8 elderly participants (56 to 75 years old). The experiment is still being performed, but we aim to finish this before summer 2022.
|
現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
The large-scale experiment with 75 planned participants has been completed basically for 2 age groups. Under the present circumstances, it is not easy to recruit elderly participants (56 to 75 years old), but we nevertheless aim to finish this experiment before summer and present/publish the results. The planning for an additional experiment has been done, and preparations are being made to perform it (see below).
|
今後の研究の推進方策 |
Following the large-scale experiment, two additional ones are planned. The first is a free dwell-time experiment, in which users are asked to select a series of objects by gazing at each object for as long as they think is appropriate. In the experiments described above, object selection took place under dwell times that were set by the experimenter. Although a dwell time of 600 ms was generally preferred by many users, there were individual differences. Personal preferences mainly concern a trade-off between speed of object selection versus certainty of object selection, related to individual goals. By varying task requirements, we aim to investigate how task goals influence the dwell time voluntarily used by participants. In a following experiment, we investigate the effect of input practice. The ultimate goal is to get complete insight into people's dwell time preferences for practical application in eye tracking systems.
|