Support for ECVP2007 abstract

Support for ECVP2007 abstract:

Whiter than white, blacker than black – overshoots in lightness perception

J Geier, L Séra, M F Hudák

All rights reserved, http://www.geier.hu/ECVP2007

Image programmer: János Geier, HTML and Flash programmer: Mariann Hudák.

János Geier janos@geier.hu
Stereo Vision Ltd,
Nádasdy Kálmán utca 34, H1048 Budapest, Hungary

László Séra lsera@uranos.kodolanyi.hu
Department of Psychology, Kodolanyi Janos University College,
Fürdő u. 1, H 8000 Székesfehérvár, Hungary

Mariann F. Hudák hudakmariann@gmail.com
Institute of Psychology, Eötvös Loránd University,
Izabella u. 46, H1064 Budapest, Hungary

Expanded abstract

We investigate a new illusion: if the contrast of a patch (Gauss patch, sharp edged disk, Gabor patch etc.) in a homogeneous background is continuously reduced until zero and the observer is fixating a signed point, then the illusory inverse of the patch is perceived at (or even before) the end of that process, or, in case of Gabor patch, an illusory displacement is perceived. The illusion occurs even on a totally white or black background, so that the inverse of a black patch on a white background will appear even brighter than the white background itself, though the dark patch is still present at that point.

The phenomenon cannot be considered as classical afterimage, since (i) it appears even at black or white patches, (ii) it appears also at short exposition (<1s), (iii) it is also perceived at low contrast (<10%), and (iv) it is eliminated by eye movements.

In our experiment (n=20) subjects had to stop the computer controlled contrast-reduction the moment when they perceived the disappearance (the beginning of turning into inverse) of the patch. Result: the patch turns into illusory inverse much before its physical contrast reaches zero.

We have modelled the phenomenon by a retinotopical set of PID controllers: according to our hypothesis, the cause of the overshoot is the 'maladjusted' D component.

According to our theoretical model, the mode of operation of each retinotopical point can be considered as a PID controller, which predicts the perceived brightness of the given point over time. When the patch merges into the background, the local lightness-alternation stops, but the maladjusted D component has proceeded much forward with the prediction of the perceived brightness by that time, so that is why a much lighter patch is seen, even if the dark patch is still present.

The explanation is the same in the case of a white patch darkening until its contrast is zero. In this case, it is the darkening of the patch what is over predicted by the D component. The illusory movement of the Gabor patch is modelled by the same concept: the dark phases of the Gabor patch overshoot into white, while the light phases overshoot into black. By the illusory replacement of each phase, the displacement of the whole Gabor-patch is perceived.

Our model also provides a new theoretical explanation for the Breathing Light illusion. By moving forward and backward, the lightness alternates at each retinal point. The retinal PID controllers predict this local brightness change forward, resulting in the illusory enlargement of the image when moving towards it.

Our model is a theoretical model, i.e. it does not refer to any concrete brain area or neural structure. The connection between our theoretical model and the knowledge about the visual system needs further investigation.

. . . > > > Abstract in Perception

Here you can see our basic phenomenon:
at the end of the contrast-decreasing process,
you will see the illusory inverse of the patch

Press the 'play' button... (To repeat: press 'play' twice.)
Fixate on an arbitrary point of the image...
(It can be either the centre, or the corners etc. )
Do you see a white patch....?
If you do, then you are a victim of an illusion :)