A Reliable Technique for Personal Identification or Verification

Abstract

Biometric characteristics are not unchangeable. Most biometric characteristics remain constant after they are fully developed. The hopes on biometric systems are in many respects justified and their use frequently gives good results. Fingerprint based biometric systems offer positive identification with a very high degree of confidence and compact solid state fingerprint sensors can be embedded in various systems. Fingerprinting does not require a laboratory for analysis, and fingerprints remain relatively constant over time, with the exception of injury. Each person has ten fingers, ten unique tokens tied to his or her identity. In this thesis, our objective is to design a fingerprint-based biometric system which is capable of achieving a fully automatic "personal identification" with a high level of confidence based on fuzzy sets in real time. Primarily the ends, branches of ridges and valleys of fingerprint are scanned. These points are unique for every person and are used to match fingerprints by their location. The following topics are covered: introduction, image requirements in acquisition and specifications, image noise reduction and enhancement, image processing methods, matching procedures and the conclusion.