Abstract
A single biological object, such as a hand, can assume multiple, very different shapes, due to the articulation of its parts. Yet we are able to recognize all of these shapes as examples of the same object. How is this invariance to pose achieved? Here, we present evidence that the visual system maintains a model of object transformation that is based on rigid, convex parts articulating at extrema of negative curvature, i.e., part boundaries. We generated pairs of jagged polygonal contours that were identical except for a single ‘part’ (an extruding segment of the contour) that pivoted across the two contours. The two contours - the ‘base’ and ‘transform’ - were both used to divide a polygonal shape into two halves resulting in four closed shapes. Similar to an ‘Attneave egg’ (Attneave, 1971), the figure and ground of the dividing contours was reversed across the two halves of the split shapes; under one figural assignment the difference between the ‘base’ and ‘transform’ contours corresponded to a convex region pivoting at concavities - i.e., a legitimate articulation - while under the other assignment it consisted of a concave ‘hole’ pivoting at convexities. In a 2AFC similarity task, subjects were much more likely to match the base shape to a transform when it corresponded to a legitimate articulation then when it didn't, even though the underlying contour transformations were identical in the two cases. These results suggest that the visual system maintains expectations about the way objects will transform, based on their static geometry.
Funded by NGA Award #HM1582-04-C-0051