Fingerprint Identification Technology of Fingerprint Scanner

The fingerprint identification of the Fingerprint Scanner, the user only needs to input the fingerprint, not the user name, the feature extraction module extracts the features of the input fingerprint, and the system repeatedly reads the fingerprint features and the input fingerprint features from the database according to a certain strategy. Match until a fingerprint that can successfully match the input fingerprint is found, then output the corresponding user name, or until no fingerprint matching the input fingerprint can be found in the database, the fingerprint will be rejected, and the fingerprint identification process will take one or more times. Matching, according to the type of identification stage, automatic fingerprint rj system is divided into fingerprint verification system (or called 1:1 system) and fingerprint identification system.

The performance parameters for evaluating the automatic fingerprint identification system are divided into three categories: correctness parameters, time parameters and space parameters. The commonly used correctness parameters mainly include error matching rate, error mismatch string, equal error probability, etc. The time parameter describes the speed of the system , such as registration time, matching time, etc. The spatial parameters mainly refer to the allocated memory size and template feature size/J. When the system matches two fingerprints, they will get their matching scores, which can be regarded as similarity. Without losing a glue, the similarity can be further reduced to [o, U, the closer the similarity is to I, the more similar the two fingerprints are, and the closer the similarity is to o, the greater the difference between the two fingerprints. If given A similarity threshold f, if the matching score is greater than or equal to ', then the two fingerprints are judged to be matched (from the same finger), otherwise the two fingerprints are judged to be mismatched (from different fingers), in this way, the given similarity reading value f , then two error rates can be obtained, namely the false matching rate and the false mismatching rate:
(1) False matching string fMR, Qinshuai M: Match the fingerprints of two different fingers, and the system considers the probability that the two fingerprints come from the same finger.
(2) Error mismatch rate MW, MNM-hehhh (or called error rejection, MebJ, tR, ): Matching two different fingerprint samples from the same finger, the system thinks that the two fingerprint samples are from different fingers , mm and 5NMR are functions of ) will become larger, when 6 decreases, the opposite is true, because Dm: f(t), so c can be expressed as '={(fMR) by FM, so delete R: 9(*)=8(F(delete) )=h(FM Min), then shun is a function of MB.
If the horizontal coordinate is taken as the horizontal coordinate, and the vertical coordinate is taken as the vertical coordinate, a plane coordinate system is established. When c takes different values, we connect the points in the coordinate system < MR(') and delete R(c)) into a curve , this curve is called Roc (RQe5veopM6qCu negative) curve, the closer the Roc curve is to the horizontal axis, the higher the correctness of the system.