Purchase this article with an account.
Jan Koenderink, Andrea van Doorn, Sylvia Pont, Whitman Richards; Gestalt and translucency. Journal of Vision 2007;7(9):910. doi: 10.1167/7.9.910.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We reconsider the stimulus configuration due to Metelli, in which (in the “translucent” experience) an observer sees two adjacent panels with a smaller, superimposed and translucent overlay. The stimulus configuration consists of four areas A, B, P and Q, such that the pairs AB, PQ, AP and QB meet at common borders, whereas the quadruple of the regions A, B, Q and P occurs in two X-junctions. Common experiences are (1) a “mosaic” in which the four areas appear as opaque and juxtaposed and (2) a mosaic of two adjacent panels A and B, overlaid with a single, translucent area W=P+Q (say), such that P is “due” to A seen through W and Q “due” to B seen through W. Observers experience difficulties with the dichotomy because they frequently “see” Gestalts differing from these two. An analysis reveals that five distinct categories may be expected. Observers experienced no difficulties with these categories. Their responses closely satisfy Metelli's constraints (which are very similar to the simplest physical model on the basis of superimposed turbid layers). Observer confusions can be understood on the basis of volatility of the constraints. The main deviation from the simplest model is that observers show an idiosyncratic preference for certain Gestalts over others, in particular, they prefer objects over holes.
This PDF is available to Subscribers Only