
A
P P E N D I X
A
–
Q
U A L I T Y A N D
C
O N T E N T
Page 52
© 1999 Biometric Identification, Inc.
Section A.3 - Using Content and Quality during Enrollments
As described in section A.1, Quality and Content scores are returned in the enrollment
process. These scores give an indication of the performance of the template enrolled. To a
large degree, the verification algorithm compensates for deficiencies in image quality and
loss of information content. Nonetheless, knowledge of these parameters and what they
mean helps ensure optimal performance of the Veriprox.
False Acceptance and False Rejection
In order to understand the effects of poor image quality
and poor information content
it is necessary to understand how to measure performance. Performance of the
Veriprox is presented in terms of False Rejection and False Acceptance.
False Rejection indicates that the Veriprox incorrectly rejected a fingerprint
that corresponds to the person’s template. False Rejections rarely occur
and primarily result from the inability to get a good image of the finger.
False Acceptance indicates that the Veriprox accepted a fingerprint that
does not correspond to the template it was compared against. False
Acceptances also are rare and primarily result when a fingerprint template
is characterized by low information in the enrolled print.
The algorithm on the Veriprox has been tuned so that the false acceptance and false
rejection rates are equal at the medium security level (level 3), delivering the industry
leading accuracy. This is known as the Equal Error Rate. Increasing the security
(e.g., changing the security level from 3 to 1) will decrease the chance for false
acceptance at the expense of increased false rejection. Reducing the security (e.g.,
changing the security level from 3 to 5) will decrease the chance of a false rejection
at the expense of false acceptance. The table below indicates the expected error
rates at the different security levels.
Security Level
False Rejection Rate
False Acceptance Rate
Very Low (5)
1 / 10,000
1 / 100
Low (4)
1 / 5000
1 / 200
Medium (3)
1 / 1000
1 / 1000
High (2)
1 / 200
1 / 5000
Very High (1)
1 / 100
1 / 20,000