Biometric systems use unique and measurable physical, biological, or behavioral traits of people to establish or to verify their identification. Biometrics are most often used to perform identity verification for authorized access to computer networks or secure facilities. The physical attributes typically used include fingerprints, face, iris, hand geometry, handwriting, or voice. Compared to common identification methods, such as identification (ID) cards, personal identification numbers (PINs) or passwords, biometrics can be more convenient for users, have lower costs for businesses, can reduce fraud, and are more secure. However, challenges for using biometrics include issues of policy and privacy.
In this presentation, we review the more commonly used biometric signals - fingerprints, face images, and iris images - and discuss the general structure of the identification algorithms. We also define some of the metrics used in evaluating biometric systems. Some of the issues that need to be considered in implementing a biometric system are also discussed. Improvements in the technology have made biometric systems more accurate, more convenient and more secure than the more widely accepted means of identification such as ID cards. In an age where identity theft is becoming more prevalent and terrorist threats substantiate the need for automatic identification of people from a distance to identify potential terrorists, biometrics appears to be an important part of solutions to these problems.