REAL-TIME MATCHING SYSTEM FOR LARGE FINGERPRINT DATABASES

With the current rapid growth in multimedia technology, there is an im minent need for efficient techniques to search and query large image d atabases. Because of their unique and peculiar needs, image databases cannot be treated in a similar fashion to other types of digital libra ries. The contextual dependencies present in images, and the complex n ature of two-dimensional image data make the representation issues mor e difficult for image databases. An invariant representation of an ima ge is still an open research issue. For these reasons, it is difficult to find a universal content-based retrieval technique. Current approa ches based on shape, texture, and color for indexing image databases h ave met with limited success. Further, these techniques have not been adequately tested in the presence of noise and distortions. A given ap plication domain offers stronger constraints for improving the retriev al performance. Fingerprint databases are characterized by their large size as well as noisy and distorted query images. Distortions are ver y common in fingerprint images due to elasticity of the skin. In this paper, a method of indexing large fingerprint image databases is prese nted. The approach integrates a number of domain-specific high-level f eatures such as pattern class and ridge density at higher levels of th e search. At the lowest level, it incorporates elastic structural feat ure-based matching for indexing the database. With a multilevel indexi ng approach, we have been able to reduce the search space. The search engine has also been implemented on Splash 2-a field programmable gate array (FPGA)-based array processor to obtain near-ASIC level speed of matching. Our approach has been tested on a locally collected test da ta and on NIST-9, a large fingerprint database available in the public domain.