The effect of edge strength on object recognition from edge images

Sanocki et al. (see J. of Experimental Psychology: Human Perception and Performance, vol.24, no.1, p.340-49, 1998) found that human competence at object recognition from viewing binary edge maps is substantially worse than that from viewing color (or gray scale) images. This article investigates a refined coding of the edge map which incorporates information about edge strength (contrast). It is shown that coding edge maps for edge contrast does improve the recognition performance. However, there is still a substantial gap in recognition performance between strength-coded edge maps and original gray scale images.