Sorting it out: Machine Learning and Fingerprints

A. Schwaighofer

Abstract:

Machine learning concepts can find applications in many domains. We describe here one such application to the problem of fingerprint verification. In biometric verification using fingerprints, one often has to handle large archives of fingerprint images. When verifying an individual fingerprint, it is important to have a fast method at hand that selects suitable candidate images out of a large archive. We present an approach to sort a fingerprint archive according to similarity with a given fingerprint, inspired by an opto-electronic device called the wedge-ring-detector.

Reference: A. Schwaighofer. Sorting it out: Machine learning and fingerprints. Telematik, 8(1):18-20, 2002.