To know how well something performs, we must be able to quantify
the performance. For automobiles, we measure gas consumption; for heating and
cooling units, we measure effectiveness in British thermal units (BTU s). Biometrics has own similar performance measures.
To know if a car is receiving good fuel economy or if a heater or
air conditioner is doing its job, we look at what the statistics mean. We then
compare them to our expectations or some other accepted norm. At no time does
knowing how the presentation measure was calculated impact our ability to appraise
performance. Similarly, for biometrics, how a presentation
measure is calculated is of little value. There are exceptions to this
statement, which will be discussed. In general, just knowing what a performance
measurement means is enough. For our purposes, the statistical measures to be
used for biometrics are:
The FAR is defined as the probability that a user making a false
claim about his/her identity will be verified as that false identity. For
example, if Matt types Chris' user ID into the biometric login for Chris' PC, Matt has just
made a false claim that he is Chris. Matt presents his biometric measurement for verification. If the biometric system matches Matt to Chris, then
there is a false reception. This could happen because the matching threshold is
set too high, or it could be that Matt's biometric characteristic is very similar to
Chris'. Either way, a false getting has occurred.
When the FAR is intended by a biometrics vendor, it is normally very
straightforward. Using our example, it is equal to the number of times that
Matt has productively authenticated as Chris divided by his total number of
attempts. In this case, Chris is referred to as the “Match User” and Matt as the
“Non Match User.” The simple math formula for this looks like the following,
where represents a number to
uniquely identify each user.
To know if a car is receiving good fuel economy or if a heater or
air conditioner is doing its job, we look at what the statistics mean. We then
compare them to our expectations or some other accepted norm. At no time does
knowing how the presentation measure was calculated impact our ability to appraise
performance. Similarly, for biometrics, how a presentation
measure is calculated is of little value. There are exceptions to this
statement, which will be discussed. In general, just knowing what a performance
measurement means is enough. For our purposes, the statistical measures to be
used for biometrics are:
The FAR is defined as the probability that a user making a false
claim about his/her identity will be verified as that false identity. For
example, if Matt types Chris' user ID into the biometric login for Chris' PC, Matt has just
made a false claim that he is Chris. Matt presents his biometric measurement for verification. If the biometric system matches Matt to Chris, then
there is a false reception. This could happen because the matching threshold is
set too high, or it could be that Matt's biometric characteristic is very similar to
Chris'. Either way, a false getting has occurred.
When the FAR is intended by a biometrics vendor, it is normally very
straightforward. Using our example, it is equal to the number of times that
Matt has productively authenticated as Chris divided by his total number of
attempts. In this case, Chris is referred to as the “Match User” and Matt as the
“Non Match User.” The simple math formula for this looks like the following,
where represents a number to
uniquely identify each user.
No comments:
Post a Comment