Purchase this article with an account.
Martin Banks, Bjorn Vlaskamp; The Venetian-blind effect: A prior for zero slant or zero disparity?. Journal of Vision 2009;9(8):45. doi: 10.1167/9.8.45.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
When sensory information is uncertain, prior expectations affect the percept. For example, perceived surface slant tends toward zero when the input signals are unreliable. This is Gogel's equidistance tendency, or a prior for zero slant. When the images presented to the two eyes differ greatly, the ability to estimate disparity worsens. This is a consequence of using interocular correlation to estimate disparity and can be thought of as a preference for zero disparity. We investigated whether a zero-slant prior or zero-disparity preference dictates the perceived slant of ambiguous stereo-defined surfaces. We presented vertical sinewave gratings of different spatial frequencies to the two eyes. As the frequency ratio differed increasingly from 1, the surface no longer appeared like a slanted plane, but rather looked like a series of small slanted patches: this is the well-known Venetian-blind effect. The zero-slant prior predicts that phantom blinds will appear when the specified slant differs too much from zero. The zero-disparity preference predicts that they will appear when the frequency ratio differs too much from 1. Both hypotheses predict the effect, including the number of blinds. To distinguish the hypotheses, we presented the stimuli in eccentric gaze where a frequency ratio of 1 does not correspond to zero slant. The data clearly showed that frequency ratio, not slant, determines the occurrence of the blinds. However, the perceived slant of the envelope of the blinds fell in-between the values predicted by the two hypotheses. Thus, the Venetian-blind effect is caused by a preference for zero disparity, but the perceived slant of the stimulus is affected by a prior for zero slant.
This PDF is available to Subscribers Only