Depth-of-Focus Affects 3D Perception in Stereoscopic Displays

Abstract

Stereoscopic systems present binocular images on planar surface at a fixed distance. They induce cues to flatness, indicating that images are presented on a unique surface and specifying the relative depth of that surface. The center of interest of this study is on a second problem, arising when a 3D object distance differs from the display distance. As binocular disparity must be scaled using an estimate of viewing distance, object depth can thus be affected through disparity scaling. Two previous experiments revealed that stereoscopic displays can affect depth perception due to conflicting accommodation and vergence cues at near distances. In this study, depth perception is evaluated for farther accommodation and vergence distances using a commercially available 3D TV. In Experiment I, we evaluated depth perception of 3D stimuli at different vergence distances for a large pool of participants. We observed a strong effect of vergence distance that was bigger for younger than for older participants, suggesting that the effect of accommodation was reduced in participants with emerging presbyopia. In Experiment 2, we extended 3D estimations by varying both the accommodation and vergence distances. We also tested the hypothesis that setting accommodation open loop by constricting pupil size could decrease the contribution of focus cues to perceived distance. We found that the depth constancy was affected by accommodation and vergence distances and that the accommodation distance effect was reduced with a larger depth-of-focus. We discuss these results with regard to the effectiveness of focus cues as a distance signal. Overall, these results highlight the importance of appropriate focus cues in stereoscopic displays at intermediate viewing distances.

(Disparity-Driven) Accommodation Response Contributes to Perceived Depth

“…Thus, to judge the depth of our stimulus, observers should have scaled the binocular disparity in the stimulus using an estimate of its egocentric distance (Bradshaw et al, 1996). This indirect effect on depth perception has already been observed in studies on the effect of focus cues on depth perception (Watt et al, 2005; Hoffman et al, 2008; Vienne et al, 2015). Our results thus suggest that bias in perceived depth indirectly ensues from undershoots in the disparity-driven accommodation response.…”

“…More recent studies have suggested that both accommodation and blur may provide a more useful cue to depth than expected (Watt et al, 2005; Held et al, 2012). Some experimental studies reported reduced depth constancy when the stimulus was displayed with conflicting accommodation and vergence distances, such that depth perception was biased toward the screen plane distance (Watt et al, 2005; Hoffman et al, 2008; Vienne et al, 2015). The reported bias could not be attributed to the accommodative response (i.e., extra-retinal signal) or the accommodative stimulus (i.e., retinal blur) because accommodation responses were not measured.…”

The Role of Vertical Disparity in Distance and Depth Perception as Revealed by Different Stereo-Camera Configurations

“…However, it is conceivable that the mean perceived distance was not affected because subjects did not rely on the same cues to estimate distance. For example, some participants might assign less weight to the vergence cue in favor of others (e.g., the accommodation or focus cue which signals screen distance, Vienne, Blonde´, & Mamassian, 2015), and thus, they perceive less depth between the object and the screen.…”

“…To help in the task, we displayed a transparent cylinder, so that observers could use the front and back face of the cylinder to better estimate its depth (Vienne et al., 2015). In this experiment, we used the method of constant stimuli to vary the depth ratio of the cylinder (i.e., the cylindricality).…”

“…To investigate how perceived depth is influenced by the vertical disparity setting, observers were required to estimate the depth of a cylinder (i.e., to indicate which, of the depth or the width of the cylinder, was larger), which was fixed at reachable distance. To help in the task, we displayed a transparent cylinder, so that observers could use the front and back face of the cylinder to better estimate its depth (Vienne et al., 2015). In this experiment, we used the method of constant stimuli to vary the depth ratio of the cylinder (i.e., the cylindricality).…”