Abstract
Victor Klymenko and Naomi Weisstein first described the Motion-Induced Contour (MIC) in 1980. In their original demo, a tilted wire cube is shown with one of its vertical edges removed. When the cube is set in motion, rotating in depth back and forth about its vertical axis, an illusory contour is perceived in the position of that missing edge. These displays are interesting because the illusory contours can be nearly orthogonal to the inducing, real edges, unlike most illusory contour displays. Although related to structure-from-motion and stereokinetic contours, MICs have several important differences. Through a series of papers, Klymenko and Weisstein concluded that the perception of the contour depends on a 3-D surface interpretation: structure-from-motion, a rigidity constraint, and is not simply due to rotation or motion in depth but rather 3D rotation through depth. Inspired by their work, we have for the first time in nearly 40 years recreated their original stimuli and have created several novel variants of their original illusion that challenge the initial conclusions and place new constraints on current models of illusory contour formation. We show that (1) unlike other types of illusory contours, MICs are minimally affected by scene context. This allows illusory contours to be observed in 'empty-space' between structure-from-motion induced surfaces. (2) Only the motion of the real contours needs to be rigid – non-rigid MICs can be formed. (3) Some MIC displays are bi-stable, so that perceived global motion changes as a function of perceived depth arrangement of the real edges, and this in turn can determine whether or not MIC's are observed, and (4) minimal displays with only two or four edges also produce illusory surfaces, which are sometimes attached to only a single real contour.
Meeting abstract presented at VSS 2018