Biometrics accuracy: Fingerprint matching using Self-Organizing Map

One of the priority when designing systems is the security. There are several means of ensuring security within a system. Biometrics is becoming widely used. Biometrics exploits the fact that there is some information that is unique to each individual, so a person can be uniquely identify within the system. We will focus in this work, on the fingerprint. Fingerprint, by nature is unique to each individual, but some errors can occurred during the processing, since it is processed as a digital image and transmit through a channel for the storage in the memory, so along these transactions noise can be introduced and distort the original images. Our system, should avoid false positive and error matching. The match should be done correctly, a person should not have the access denied while his fingerprint is stored in the database, or a fingerprint of a person stored in the database should not match with the one for another person either in the database or an intruder. Self Organizing Map (SOM) has some excellent features that can be used for fingerprint identification. In our work we are going to design a SOM for fingerprint identification.