"Five Illusions Challenge Our Understanding of Visual Experience"
Vision Sciences Society (2025) [Poster] [Abstract]

"Five Illusions Challenge Our Understanding of Visual Experience"
Vision Sciences Society Demo Night (2025) [Link]

The Hollow Face Illusion is often appealed to as evidence that visual experience is the visual system's "inference" ("best guess") about what is out there in the world, since a hollow mask is percieved as inverted in depth.

However, in two illusions I question whether percieved depth is really inverted:

I show that if you add physical objects into the hollow of the Hollow Face Illusion - space that physically exists, but which is "impossible" according to the illusion - the illusion still persists (as evidenced by illusory motion), but you still see the true (non-inverted) depth ordering of the objects and mask in stereo depth.

Kriegeskorte Lab experiencing the Un-Hollow Face Illusion.

I show that if you add balls to the tip of the nose and the cheek, and morph back and forth from a receding to a protruding mask, the change in the ordinal depth of the balls (inverting back and forth) is apparent.

If the Un-Hollow Face Illusion suggests that 3D Shape is governed solely by stereo vision, then the next question is how the visual system extracts 3D shape from disparity. Traditionally, it is thought that retinal disparities are "scaled" to transform them from retinal to world coordinates.

However, I argue there is no "depth constancy" (scaling into world coordinates):

Linton, P. (2024). Linton Stereo Illusion. Preprint

Traditional accounts ask participants to make evaluative depth judgments. I simplify the question by testing which of two conditions appears to move rigidly in depth (with angular size controlled): a stimulus moving in depth with contant physical separation, or a stimulus moving in depth with contant disparity?

As you see below, the answer is "constant disparity" (right), suggesting that percieved stereo depth is simply a function of disparities on the retina.

Presented at Demo Night at the Vision Sciences Society (VSS)(2024) and European Conference on Visual Perception (ECVP)(2024).

We know that increasing the separation between the eyes makes scenes seem miniature. Traditionally, this has been explained as due to "vergence" (eye rotation) making objects seems smaller and closer.

However, I argue that it instead rests on a purely cognitive association between accentuated stereo depth (from horizontal disparities) and closer distances given disparities fall off with distance squared:

Linton, P. (2024). Visual Scale is Governed by Horizontal Disparities: Linton Scale Illusion. Applied Vision Association 2024 Christmas Meeting Abstract

I replicate the effect in VR (Condition 1), but show it disappears as soon as we control for changes in horizontal disparities (Condition 2).

Switching between the two conditions we see a startling effect: the "scale" of the scene changes, even though all that has changed is the geometry of the scene, not the apparent distance of the scene itself.

Size constancy is thought to change the percieved angular size of objects in images, for instance the apparent size of the cars in this image:

However, I argue that size constancy does not affect perceived angular size:

Linton, P. (2025). Size and Shape Constancies Do Not Affect Perceived Angular Size: Linton Size Constancy and Shape Constancy Illusions. Applied Vision Association 2025 Spring Meeting Abstract

If we add frames to the cars, the cars appear more distorted than the frames, which is inconsistent with size constancy affecting perceived angular size.

Color constancy is thought to affect the percieved color of objects, so that the disk in this illusion by Akiyoshi Kitaoka (based on Anderson & Winawer, 2005) is perceived differently as yellow or blue:

However, I argue that color constancy does not affect perceptual appearance:

Linton, P. (2025). Lightness and Color Constancies Do Not Affect Perceptual Appearance: Linton Lightness and Color Constancy Illusions. Applied Vision Association 2025 Spring Meeting Abstract

I show that if we switch back and forth between the two interpretations of the disks, our judgement of the disk's color ("yellow" or "blue") changes without the disk's perceptual appearance changing.

© Paul Linton, 2025. Site based on Jon Barron's website.