Abstract
Purpose:
Nonfixated laterally drifting blurred targets appear sharper than when they are stationary. The purpose of this study is to determine whether a nonfixated blurred image moving in stereoscopic depth would appear similarly sharpened.
Methods:
15 normally binocular young adults viewed, in a mirror stereoscope, digital images 40cm from the eyes. The 3D images were two 1°w x 2°h bars of 14.5 cd/m2 luminance and 50% contrast on a dark gray background. The bars appeared simultaneously for 0.5 sec, above and below a 5 minarc white fixation spot. The moving bar disparity changed linearly at velocities of 1.75, 3.5, or 7Δ/sec while the other bar was static. The direction of disparity change (crossed or uncrossed) and moving bar placement (above or below fixation) were randomized between trials. The edges of the bars were blurred by a cosine luminance profile, which was fixed at 50 minarc for the moving bar and adjusted by a QUEST adaptive staircase for the static bar to match the moving bar blur. In 2D control trials the bars were presented in a similar manner except that the moving bar motion was lateral rather than in depth. All conditions were repeated 3 times.
Results:
A mixed ANOVA evaluated the effects of motion type (2D vs. 3D), subject, and velocity (1.75, 3.5, or 7) on match blur. Match blur decreased significantly with velocity (f=149, p=0.0), revealing significant neural sharpening. There was no difference of sharpening between 2D and 3D motion (f=0.26, p=0.77), demonstrating that neural sharpening occurs in stereoscopic motion-in-depth. At the highest velocity, the moving bar edges were sharpened by an average of 19% compared to the lowest velocity. Neural sharpening differed significantly between subjects (f=2.2, p=0.002), ranging from 8% to 36%.
Conclusions:
This is the first report to show that neural sharpening occurs in stereoscopic motion-in-depth. Neural sharpening in stereoscopic motion is twice as strong as lateral motion sharpening when referenced to retinal image velocity, but when expressed in cyclopean angular units (i.e., 1Δ disparity = 1Δ of lateral motion), neural sharpening is equally strong in stereoscopic and lateral motion. Our results suggest that nonfixated optically blurred stereoscopic images moving in depth in 3D displays appear clearer than predicted from retinal image focus.
Keywords: 434 binocular vision/stereopsis •
601 motion-2D •
602 motion-3D