3-D Shape Pop-out

Our experiments show that when one upside-down shaded cube is embedded among multiple upright shaded cubes, the task can be processed by early vision mechanisms.
Perceptually, the upside-down cube pops out from the others, and is detected effortlessly whether it is surrounded by few or many upright cubes.

One question that comes that mind is whether the shading is actually important. We test this using line cubes instead of shaded cubes:

Our experimental results show that the upside-down line cube is not easy to detect, requiring a processing time that increases as the number of surrounding upright line cubes is increased. Would you agree that the perceptual pop-out effect seen with the shaded cubes is missing in this line display?

Another intriguing result we found is that, while an upside-down shaded cube among upright ones pops out, the converse is not true:

In this display, the odd-man-out is the upright cube, surrounded by upside-down ones. Did it pop-out at you, or did you have to search for it?

This upside-down cubes display is actually the exact 180-degree rotation of the upright cubes display shown at the top of this page. Since these two displays are exactly the same in 2-D terms, such a strong perceptual difference points to the involvement of 3-D shape processes.

What do you think this perceptual difference is due to? We think it has to do with the relationship between 3-D shape and reflectance.

Reference:

Sun, J.Y. & Perona, P. (1996). Preattentive perception of elementary three-dimensional shapes. Vision Research, 36(16), 2515-2529.