A NEW METHOD OF ESTIMATING SHAPE SIMILARITY

The goal of this paper is to describe a measure of similarity among sh apes which ''look'' similar, such as the same letter in different font s, or pictures of different, yet ''perceptually'' similar, chairs. We represent shapes as special weighted graphs, the vertices of which rep resent the ''lumps'' of the shapes in a given orientation. We then red uce these graphs using a ''small leaf'' trimming procedure until the r esulting graphs are isomorphic. A similarity measure is calculated bas ed on this representation via a polynomial-time complexity algorithm. A significant improvement in the complexity of the method presented is provided by eliminating the need to minimize the similarity measure o ver all orientations. A test database on which the method has been exa mined consists of 25 pairs of perceptually similar shapes, which inclu des the letters of the Hebrew alphabet.