more typically rounded, blunted, blurred, ragged, textured, etc. Some of these attributes are listed and described in A failing of many gray-level corner detectors is that they do not extract most of the attributes of a corner apart from its more detail below. In particular, cornerity is usually destrength. This paper provides several postprocessing techniques rived from a subset of the following properties:for determining additional corner attributes (i.e., color, orientation, subtended angle, and contrast). Corner matching processes• Position-location of the corner. For round corners can use this additional information to resolve otherwise ambigu-the two edges of the corner can be extrapolated and the ous correspondences and to eliminate corners whose attributes position defined by their intersection. Alternatively, the do not match certain criteria.