In order to protect users of computer systems and to secure network-based transactions, demand is increasing for improved user authentication procedures to establish the identity of an actual user and to bar access to a terminal to anyone who is unauthorized. Personal identification using biometrics, i.e., a person’s physical or behavioral characteristics has come to attract increased attention as a possible solution to this issue and one that might offer reliable systems at a reasonable cost. While traditionally this technology has been available only with such expensive, high-end systems as those used in law enforcement and other government applications, today many personal–level applications have also become possible thanks to the advancements in pattern recognition technology.
When compared with the conventional authentication methods that are based on “what only the person possesses” or “what only the person knows”, biometrics authentication offers two distinctive advantages:
➢ Enhanced convenience: By merely presenting this biometrics features, a user can easily prove himself or herself. There are no troubles such that authorized users are denied access because of loss of a card or forgetting a password.
➢ Augmented security: The reliable rejection of imposters, who might attempt to gain access either by stealing or forging cards or by guessing or fraudulently obtaining passwords, becomes possible.
Among various modalities in biometrics such as fingerprints, face, iris, etc., fingerprints are the most widely used and have the longest history in real-world law enforcement applications. Research into automated fingerprint identification began in the 1960’s and the resulting AFIS (Automated Fingerprint Identification Systems) have been used worldwide with established dependability. Millions of identifications over a century of actual forensic history have clearly shown that fingerprints are unique and permanent and thus that the fingerprint identification is extremely reliable. Recent technical advances have made identification (i.e., one –to-many matching) systems low enough in cost for civilian applications.
Fingerprints have among many, the following two advantages when compared with other modalities:
1) Stable, reliable and highly accurate identification software is currently available even for use on personal computers. 2) Fingerprint sensors can be made small and thin enough to be implemented easily on small computers and even on pocket-sized terminals.
2. FINGERPRINT-BASED PERSONAL AUTHENTICATION:
A Fingerprint-based personal authentication system operates in two distinct modes: enrollment and authentication (identification), as is shown in Fig. 1. During enrollment, a fingerprint image is acquired from a finger presented by an authorized user using a “fingerprint sensor” and relevant features are extracted by the features extractor. The set of extracted features, also referred to as a “template” is stored in a database, along with the user’s information necessary for granting service and some form of ID assigned for the user.
When the user seeks for a service, i.e. in authentication mode, the user inputs his assigned ID and presents his fingerprint to the sensor, The system captures the image, extracts (inputs) features from it and attempts to match the input features to the template features corresponding to the subject’s ID in the system database. If the calculated similarity score between the input and the template is larger than the predetermined threshold, the system determines that the subject is who he claims to be and offer the service, otherwise the claim.
In identification mode, on the other hand, the user who seeks for a service presents his fingerprint only without his ID and the system may either be able to determine the identity of the subject or decide the person is not enrolled in...
References:  W. F. Leung, S. H. Leung, W. H. Lau, And A. Luk,Fingerprint Recognition Using Neural Network, Neural Networks For Signal Processing – Proceedings Of The 1991 IEEE Workshop
 M. Hartman, “Compact Fingerprint Scanner Techniques” In Proc. Biometric Consortium 8th Meeting, San Jose, Ca, June 1996
 J. Schneider, “Improved Image Quality Of Live Scan Fingerprint Scanners Using Acoustic Backscatter Measurements,” In Proc. Biometric Consortium 8th Meeting, San Jose, Ca, June 1996.
 K. McCalley, D. Setalk, S. Wilson, And J. Schmitt “Improved Image Quality Of Live Scan Fingerprint Scanners Using Acoustic Backscatter Measurements,” In Proc. Biometric Consortium 8th Meeting, San Jose, Ca, June 1996.
Please join StudyMode to read the full document