Biometric Authentication      

Biometrics statistically measure certain human anatomical and physiological traits that are unique to an individual. Automation of biometrics to electronically identify user characteristics is an effective means of user authentication for the purpose of Information Systems security.  Examples of biometrics include fingerprints, retinal scans, speaker recognition, signature recognition, hand recognition and others.

Emerging in government programs and niche commercial applications, biometric authentication is being used successfully either by itself or with other authentication techniques such as a smart card, password or another biometric. Such niche applications, programs and pilots demonstrate that biometrics are safe, reliable and user-friendly. Applications show that biometrics can lead to more convenience and reduced costs by converting from manual processes. Automation of certain applications are enabled through the use of biometrics where other authentication methods do not meet the project's authentication needs.

Since September 11th the US government has put serious effort into physical access applications for government employees and border control applications amongst others.

The underpinnings of biometric authentication comprise of FOUR STEPS;

• one is to capture human input,

• two is to filter out unwanted input such as noise,

• three is to generate a statistical representation of the biometric input, and

• four is to perform a match of the biometric instance created in step three against biometric information previously gathered, processed and stored during an enrollment procedure (generally referred to as the stored biometric model, template or print).

 The two methods of biometric authentication are defined below:

• VERIFICATION is the process of verifying the user is who they claim to be. A claim of identity, such as an account number, needs to be made during speaker verification.  The matching process described in step four is One to One matching where the biometrics instance generated in step three during verification is matched to the model the user claims to be.
• IDENTIFICATION is the process of identifying the user from a set of known users.  A claim of identity does not need to be made during speaker identification.  The matching process described in step four is One to Many matching where the biometrics instance generated in step three during identification is matched to a set of many known models. 

The goal of biometric authentication is to determine if the user is the authentic enrolled user or an impostor.  Assuming the user has previously enrolled, positive verification or identification consists of quality input, processing and successful outcome of the matching process.