The role of convexity in the corner enhancement effect, in visual short-term memory, in perception of symmetry, and in shape interference

Contour curvature information has been shown to have an impact on the visual perception of shape. We have conducted studies on perception of convexity and concavity in relation to memory and attention. Previous studies (Badcock & Westheirner, 1985; Krose & Julesz, 1989; Nakayama & Mackeben, 1989) have proposed that visual space is influenced by corners. Recent studies by Cole, Burton and Gellatly (2001) found that reaction times were faster for a stimulus located in the region of a corner of a figure. Cole et al (2001) believe that the role of corners is greater than that of straight edges, due to corners receiving a higher distribution of attentional resources relative to straight edges. The first part of this thesis considers the role figure-ground plays in the corner enhancement effect. Results demonstrate that the corner enhancement effect is only found when the probe is presented on the surface that owns the corner. Thus the corner enhancement effect is present for both concave and convex vertices. However, the effect disappears when the probe lay on the surface that does not own the corner. The second series of experiments made use of a shape with multiple concave or convex features as part of a change detection task, in which only a single feature could change. The results provided no evidence to suggest that convexities are special in visual short-term memory. Though coding of convexities as well as concavities provided a small advantage over an isolated contour. This finding is in accordance with the well documented effect of closure on shape processing (Elder & Zucker, 1993). It has been reported that deviations from symmetry were easier to detect when carried by convexities compared to deviations carried by concavities (Hulleman & Olivers, 2007). We extended this investigation to shapes that were repeated instead of reflected, to test whether there is a specific convexity advantage for bilateral symmetry. The results supported a convexity advantage for repetitions but not for reflections. Possible explanations for this are discussed. The final series of experiments involved a shape interference task; observers responded to circles or square in the context of irrelevant circles and squares. The findings suggest that interference between the shapes is much stronger when the contours that define the shapes belong to the same surface. In summary, we conclude that convexity and concavity are important aspects of shape analysis and representation, but there is no basic difference in how convexities and concavities are attended to, both in the corner enhancement effect, and in visual-short term memory. However, convexity plays a role in some perceptual tasks for example, when analyzing complex shapes observers may adopt strategies that focus on the convexities.