Purchase this article with an account.
Jonathan Q. Hal, Bosco S. Tjan, Zili Liu, Choon-Kyu Lee, Arash Motamed; Tracking a stereo-kinetic ellipse. Journal of Vision 2004;4(8):253. doi: https://doi.org/10.1167/4.8.253.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
A rotating ellipse mostly leads to the percept of a circle wobbling in 3D. Only when the ellipse is highly elongated or with effort, can one perceive a 2D ellipse rotating in the image plane. We investigated the patterns of eye movement when subjects perceive either the ellipse in 2D or the circle in 3D. We hypothesize that this pattern reflects the difficulty of tracking a physical point on the perceived stimulus. A stereo-kinetic display of an ellipse spanning 22×16 deg of visual angles was viewed monocularly. The ellipse rotated continuously in the image plane with a period of 8 sec/cycle. In alternating trials, a trained subject was asked to establish and hold either the 2D or 3D percept. The subject indicated having established a percept by pressing a button. The trial continued for the next 16 sec. A head-eye tracker monitored the subject's eye movement, captured the response, and was synchronized with the display. To maintain the 2D percept, two subjects were found to track one of the 2 highest-curvature points along the ellipse's contour. In contrast, they viewed the stimulus's feature-less interior to maintain the 3D percept. A third subject seemed to attend uniformly to the interior of the stimulus regardless of his percept. Subsequent debriefing revealed that he was imagining a figure that filled the interior. While it is straightforward to explain the eye movement pattern in the 2D percept, we can also explain the pattern in the 3D percept based on a computational theory of slowest motion. When a circle wobbles in 3D, the 2D projection of any of its physical points has an image motion much slower than the 2D rotational speed of the stimulus. Moreover, the 2D curvature of this point's projection on the image changes over time, making it impossible to track. Since no physical point along the circular contour is distinct, and since they are equally difficult to track, subjects tended to view the feature-less center instead.
USC Undergraduate Research Program Grant (BT)
This PDF is available to Subscribers Only