Application areas Biometrics

In an automatic fingerprint identification system (AFIS), the fingerprint database can be huge, often tens of thousands of fingerprints. If you were supposed to check for similarity between the query fingerprint and every other fingerprint in the database, this could take very much time. With this in mind, fingerprint classification is an important step that has to be implemented in every fingerprint identification system. The classification criteria most widely used are the modification or extension of the standard Galton-Henrys classification system (Henry, 1990). Here, fingerprints are divided into 5 subclasses: whorl (W), right loop (RL), left loop (LL), arch (A) and tented arch (TA).

Pattern Solutions has developed a hybrid system for fingerprint identification based on such a regime, where one first is doing a classification stage to identify the category of fingerprint and then a total search through the actual sub database. By using such a regime one is able to reduce the query time with a factor about four.