Abstract

This paper describes a graph-matching technique for recognising line-pattern shapes in large image databases. We use a Bayesian matching algorithm that draws on edge-consistency and node attribute similarity. This information is used to determine the a posteriori probability of a query graph for each of the candidate matches in the data-base. The node featurevectors are constructed by computing normalised histograms of pairwise geometric attributes. Attribute similarity is assessedbycomputing the Bhattacharyya distance between the histograms. Recognition is realised by selecting the candidate from the data-base which has the largestaposteriori probability. 1 Introduction Broadly speaking there are two sources of information that can be tapped in the content based retrieval of images from large data-bases. The first of these is to use a compact summary of the image attributes. One of the best known examples here is the attribute histogram originally popularised bySwain and Ballard [1]...

Citations