Perceived motion transparency can overide luminance / color cues which are inconsistent with transparency (ARVO 2000 demo).

abstract (pdf)
poster (pdf)

Note: To view these demo you need a Quicktime plugin in your browser or a Shockwave capable browser.
 

When rotated rigidly about its center, the figure made of two overlapping disks splits, usually after a while the first time, and two disks are perceived to move independently sliding over each other in circular translational motion while staying upright (Wallach '76 (1)).
Click here for the Quicktime version (340 Ko). Click here for the Shockwave version.

This is what the real motion of the stimulus is in fact.
Click here for the Quicktime version (340 Ko). Click here for the Shockwave version.

In this demo, the luminance of the intersections is not consistent with physical transparency. However, transparent motion can still be perceived.
Click here for the Quicktime version (340 Ko). Click here for the Shockwave version.

 
These observations seem to contrast with those obtained with moving gratings:
Two superimposed moving gratings can be perceived either as a plaid moving rigidly or as two gratings sliding over each other (Wallach '35, '96 (2); Adelson & Movshon '82 (3)). Stoner et al. ('90, '96 (4)) showed that the perception of motion transparency was affected by whether the luminance of the gratings' intersections was consistent with physical transparency or not (see the plaid demo Stoner story, and used the back button of your browser to come back to this page).
The stimulus that we presented here is another ambiguous motion displays where globally rigid (as with the polygons) and transparent, non-rigid interpretation compete. In this display, however, the luminance cues have a weak influence on the perception of motion transparency. Moreover, segmentation based on motion can override conflicting luminance cues.

In fact both displays give contradictory results only when short exposure methods are used. When dynamic measures are used, the role of static segmentation cues can be observed in both displays: for the disks, we observed that the response time to see transparency was significantly longer when the luminance of the intersections was not consistent with transparency.
Similarly, most plaids can be perceived as sliding when observed  for a long enough time,  even when the luminance of the gratings' intersections is not consistent with physical transparency  (go to the plaid demo page).

---

(1) Wallach, H. On Perception. New York: Quadrangle, 1976, 490 p.

(2) Wallach, H. Uber visuell wahrgenommene Bewegungsrichtung. Psychologische Forschung 20: 325-380, 1935.
     Wuerger, S., Shapley, R., and Rubin, N. "On the visually perceived direction of motion" by Hans Wallach: 60 years later. Perception 25: 1317-1368, 1996.

(3) Adelson, E.H., and Movshon, J.A. Phenomenal coherence of moving visual patterns. Nature 300: 523-525, 1982.

(4) Stoner, G.R., Albright, T.D., and Ramachandran, V.S. Transparency and coherence in human motion perception. Nature 344: 153-5, 1990.
     Stoner, G.R., and Albright, T.D. The interpretation of visual motion: evidence for surface segmentation mechanisms. Vision Res. 36: 1291-310, 1996. 

---

back to home page