Comparison of gaze accuracy and precision in real-world and virtual reality

Abstract

Virtual reality (VR) is popular across many fields and is increasingly used in sports as a training tool. The reason, therefore, is recently improved display technologies, more powerful computation capacity, and lower costs of head-mounted displays for VR. As in the real-world (R), visual effects are the most important stimulus provided by VR. However, it has not been demonstrated whether the gaze behavior would achieve the same level in VR as in R. This information will be important for the development of applications or software in VR. Therefore, several tasks were designed to analyze the gaze accuracy and gaze precision using eye-tracking devices in R and VR. 21 participants conducted three eye-movement tasks in sequence: gaze at static targets, tracking a moving target, and gaze at targets at different distances. To analyze the data, an averaged distance with root mean square was calculated between the coordinates of each target and the recorded gaze points for each task. In gaze accuracy, the results showed no significant differences between R and VR in gaze at static targets (1 m distance, p > 0.05) and small significant differences at targets placed at different distances (p < 0.05), as well as large differences in tracking the moving target (p < 0.05). The precision in VR is significantly worse compared to R in all tasks with static gaze targets (p < 0.05). On the whole, this study gives a first insight into comparing foveal vision, especially gaze accuracy and precision between R and VR, and can, therefore, serve as a reference for the development of VR applications in the future.