Abstract
At last year's VSS (Wagemans et al., VSS-2001, Abstract No. 97), we presented results on priming in picture identification on the basis of contour fragments with specific curvature properties. Contours were built up gradually in 20 steps (i.e, increasing fractions of the total contour), each presented for 500 ms. One of the surprising results then was that fragments around inflections (i.e., where curvature changes from positive to negative) yielded better identification rates than fragments around curvature extrema (i.e., where positive or negative curvature is locally maximal or minimal). In the present study, we used more stimuli (N = 188), more subjects (N = 200), and a different method in an attempt to replicate and further substantiate this finding. Fragments were now centered around points along the contour which an independent sample of observers had marked as most salient (see De Winter et al., VSS-2002, submitted) or around the points midway between two popular points. Four fragmentation levels were used (15, 20, 25, or 30% of the total contour) but only one fragmented version was shown per stimulus (for 5 s each), counterbalancing stimulus identity and condition across observers. Results showed that recognition was about 7% easier for fragments centered around the midpoints than for fragments centered around the popular points, at each of the four tested percentages (45 vs. 52%, 52 vs. 59%, 57 vs. 64%, and 62 vs. 67%, respectively). This robust finding contrasts sharply with the results which we obtained in a related study, using connected straight-line versions, where the most popular points were clearly more useful than the midpoints (see Wagemans et al., VSS-2002, submitted). This suggests that identification on the basis of fragments is indeed special as suspected long time ago (see Kennedy & Domander, 1985, Perception, 14, 367–370 and a critique by Deregowski, 1986, Perception, 15, 217). More research is needed to explain this difference.
Acknowledgment: This research was supported by a grant from the Research Council at the University of Leuven (OT/00/007).