Abstract
Aspect ratio (AR) is a relative object property that denotes a particular relationship between an object's height and width (AR = H/W) when considered within an object based reference frame. The visual system's ability to extract AR information from a stimulus can promote object constancy as it is invariant under many viewing conditions. Regan and Hamstra (1992) demonstrated AR specific, size-independent, adaptation to rectangles. They proposed a possible mechanism for AR estimation that involves constructing a representation of the rectangle's diagonals and determining their angle of intersection. This implicit angle identifies the AR of the object unambiguously. To test this hypothesis, we compared the effects of adapting to rectangles, angles, and single oriented lines that correspond to a given AR. The observed shift in perceived AR is consistent with the implicit angle explanation: the direction of the shift is predicted by the AR of the adapting stimulus and the magnitude of the shift increases as the adapting stimulus configuration becomes more like the rectangular test stimuli. Rectangle adaptation yielded the greatest shift, angle adaptation produced an intermediate shift, and oriented line adaptation generated a minimal shift. Possible underlying mechanisms include orientation-tuned V1 cells (Hubel & Wiesel, 1959) and V4 cells that prefer non-Cartesian stimuli (Gallant, Braun, & Van Essen, 1993). Understanding how the visual system processes AR information will be an important link between low-level stimulus properties (e.g. orientation and spatial frequency) and high-level object identification (e.g. apples and faces).
GallantJ. L.BraunJ.EssenD. C.Van(1993). Selectivity for polar, hyperbolic, and Cartesian gratings in macaque visual cortex. Science, 259, 100–103.
HubelD. H.WieselT. N.(1959). Receptive fields of single neurons in the cat's striate cortex. Journal of Physiology, 148, 574–591.
ReganD.HamstraS.(1992). Shape discrimination and the judgement of perfect symmetry: dissociation of shape from size. Vision Research, 32, 1845–1864.