FINGERPRINT MATCHING

摘要

A method for authenticating a claimed identity, comprising acquiring a first set of data, generating a first set of filters (step Sl), associated with said first set of data, receiving a second set of filters (step S3), associated with a previously stored set of data associated with said claimed identity, for each filter in said first filter set, determining with which filter in said second set it has the greatest correlation (step S3, S4), determining a first subset of said first set of data (step S7), including data corresponding to filters in said first set of filters that have a greatest correlation exceeding a given threshold (Tl), and determining a measure of similarity (step S 14) between the first subset and a second subset of said second set of data, including data corresponding to filters in said second set of filters that have a greatest correlation exceeding said given threshold (Tl), the first and second subsets having equal length. According to this approach, correlation between the two filter sets is used to establish which data elements correspond to each other. The data sets are then reduced to form subsets containing only these corresponding elements, and therefore being equal in size. These subsets can then be used to determine a measure of similarity, for example using existing conventional measures such as Euclidian distance or ridge distance.