performance_fall05

performance_fall05 - October 4, 2005 WVU 1 Performance...

Info iconThis preview shows pages 1–6. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: October 4, 2005 WVU 1 Performance Evaluation BIOM 426 Instructor: Natalia Schmid October 4, 2005 WVU 2 Statistical Measures for Biometrics s False Accept Rate (FAR) FAR is the probability that a user making a false claim about his/her identity will be verified as that false identity. Major reasons: threshold, biometrics of claimed identity is close to biometrics of t he user making claim. FAR characterizes the strength of the matching algorithm. s False Reject Rate (FRR) FRR is the probability that a user making a true claim about his/her identity will be rejected as him/herself. Major reasons: threshold, the presented biometric features are not close enough to t he biometrics template in the database. FRR characterizes the robustness of the algorithm. s False to Enroll (FTE) FTE is the probability that a user attempting to biometrically enroll will be unable to. Rule of 3 attempts. Major reasons: the biometric hardware; an algorithm that is not tuned properly; a u ser who is slow in learning how to submit biometrics. s Equal Error Rate (EER) (Will be defined later) October 4, 2005 WVU 3 Basic System Errors s Before designing any system, biometric error rates have to be analyzed. s Two important facts: Any biometric system will make a mistake; The true value of error rates cannot be theoretically derived. s Since biometrics measurements (2D representation) or templates are available for verification of individual, a metric, S(I,T), (a measure of closeness or difference between two templates) has to be introduced. s Sources of noise: preprocessing, mapping, thermal noise, discretization, quantization, acquisition error, etc. October 4, 2005 WVU 4 Hypothesis Testing Problem s The problem of verification can be stated as a hypothesis testing problem. s Two hypothesis are possible: H0: Two samples match (null) HA: Two samples do not match (alternative) s Another statement: A and are two biometrics H0: A = (null) HA: A (alternative) s To decide between two hypothesis, the score S(I,T) is computed H0: S(I,T) > (null) HA: S(I,T) < (alternative) October 4, 2005 WVU 5 Score Distributions s The reliability of the score is influenced by Variations in the live biometrics in time; Variations from sensor to sensor; Variability in the sampling process. s Because of these reasons S(I,T) will never be one for...
View Full Document

This note was uploaded on 03/04/2012 for the course BIOM 426 taught by Professor Staff during the Fall '08 term at WVU.

Page1 / 25

performance_fall05 - October 4, 2005 WVU 1 Performance...

This preview shows document pages 1 - 6. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online