Applying algebraic and differential invariants for logo recognition

The problem of logo recognition is of great interest in the document domain, especially for document databases. By recognizing the logo we obtain semantic information about the document which may be useful in deciding whether or not to analyze the textual components. Given a logo block candidate from a document image and a logo database, we would like to determine whether the region corresponds to a logo in the database. Similarly, if we are given a logo block candidate and a document database, we wish to determine whether there are any documents in the database of similar origin. Both problems require indexing into a possibly large model space. In this contribution, we present a novel application of algebraic and differential invariants to the problem of logo recognition. By using invariants we have shape descriptors for matching that are unique and independent of the point of view. The algebraic invariants handle cases in which the whole shape of the logo is given and it is easy to describe. The differential invariants cover complex arbitrary logo shape and handle situations in which only part of the logo is recovered. We outline a hierarchical approach to logo recognition and define methods for page segmentation, feature extraction, and indexing. We demonstrate our approach and present results on a database of approximately 100 logos.