As known in biometrics: https://en.wikipedia.org/wiki/Biometrics#Performance
False match rate (FMR, also called FAR = False Accept Rate): the probability that the system incorrectly matches the input pattern to a non-matching template in the database. It measures the percent of invalid inputs that are incorrectly accepted. In case of similarity scale, if the person is an imposter in reality, but the matching score is higher than the threshold, then he is treated as genuine. This increases the FMR, which thus also depends upon the threshold value
For example we have case:
- 1000 users templates of faces in database
- 100 users-impostors which aren't in database trying to log the system, once every
- FAR = 0.001 (0.1%)
Then how many false matches will there be?
- 1000 * 100 * 0.001 = 100 (some of users-impostors will be accepted with many templates in database, some no once)
- 100 * 0.001 = 0.1 (imposters will not be allowed into the system, only 10% that one impostor may still be accepted to the system)
Or the same question in other words, what specifically is FAR (False Accept Rate):
- Is FAR a false match of one attempt to compare with only one template in database? And it does not depend on the size of the database.
- Or is FAR - if there is at least one match with any of the templates in database? And it depends on the size of the database.