How does the presence of variously oriented gradients affect the perception of depth?

Griffin Leonard

Shape can be perceived purely from intensity differences. This is called shape from shading. This illusion is designed to test how this shape perception is affected by differently orientated gratings which are not consistent with a single light source. The first image is a control. The circles top-lit circles are generally perceived as popping out of the screen, and the bottom-lit circles are generally perceived as divots into the screen. For each of the images below, please answer the following questions: (a) Do you perceive depth? If so, is the effect stronger or weaker than the control image (the first image)? (b) Is depth perceived for every circle or only certain orientations? If only certain orientations, which ones? (c) If depth is perceived, how many look circles like they go into the screen and how many look like they are coming out of the screen? (d) Are any of the images bistable? Bistable means that the perceived depth changes or oscillates between multiple interpretations.







Ramachandran, V. S. (1988). Perception of shape from shading. American Journal of Ophthalmology, 105(4), 446. doi:10.1016/0002-9394(88)90349-2

Comments


Griffin Leonard

c) The subjective reports indicate that there is a strong prior influencing perceived depth for some orientations, specifically when the white part of the gradient is on top the circle looks like a hill that pops out of the screen. The other gradients seem to be bistable for most people, however a specific depth effect is generally seen for circles with the same gradient when fixating on any particular circle. The observations answer the original question posed only partially: the perceived depth of gratings is definitely affected by nearby gratings, but the exact relationship is not totally clear.

d) To figure out exactly how orientation and spacing of the circles affects depth, I could create dozens more images to serve as intermediates between the control condition (which can be explained) and each of the experiment conditions. Using these responses I can also ask more specific questions about how the depth is perceived than the four broad questions I initially asked. For example: "When fixating on a perceived hill in the center of the image, what is the perceived depth of each of the gratings in the periphery?

Cesar Duran

a) I noticed that all the circles with a lightness coming from above were always interpreted as hills, and the circles with a lightness coming from below were always interpreted as valleys. While the left and right oriented circles were bistable, I felt it was easier to interpret them as valleys when there were grouped with top light source circles. Perhaps that the hills created by the top light sourced circles lessen the probability that the other circles would be interpreted as hills. In the final picture I notice that it isnt completely horizontal, so my interpretation for a given circle skews towards a hill/valley when it is slightly more top/bottom oriented.

b) This experiment answers the question posed. The perception of hills/valleys within the illusions is influenced by nearby gradients, especially if the nearby gradients are vertically oriented.
I forgot my name oops, Cool experiment!

Michael Anoke

For each of the images, the circles that appear to have a light source coming from above look like bumps, and the depth looks about the same as the control. For every image other than the control, it is difficult to see the "craters". They are bistable, going from a crater to a circle floating above the background. If the crater is nearby a circle that looks like a bump, then the crater is much easier to see. The other circles are also bistable, changing from a bump to a floating circle depending on where I look.

b. I believe the experiment answers the question fairly well. It might have helped to have a more controlled change between some of the images so better monitor what exactly influences the change in perception. Some images seem to have a lot of the circles changed from the previous one, so for some of the images it is difficult to tell if the grating itself doesn't coney much depth or if the surrounding grating are an influence.

Maddie C

General note: I usually have a hard time seeing divots in this illusion, bumps are much easier for me to see in general.

1- white at top looks like bumps
2- definitely multistable. for. very brief moments it feels like i see all of them as bumps
3- i only really see the ones with white at top as bumps
4- multistable, can see any set of circles as bumps (excluding when white is at the bottom).
5- similar as 4 but it seems like the sets don't seem to need to be consistent with a single if they're far away from each other? but i find it a bit hard to tell
6- same as 5

Richard McWalter

For the control I see depth, with 4 bumps and 4 divots. In the second image, they look a little bistable, but mostly like circles floating above the background. For image 3, 4, and 5, as long as I fixate on one of the circles where the light source is coming from above, the other 4 look like divots. The last image I perceive depth, even though the light is coming from different angles. Very cool!