Combining biometrics
The combination of biometric technology can enhance the accuracy of the total system. This is called multi-level biometrics (or sometimes multi-modal biometrics). Statistics however indicate one has to be very careful with combining biometrics. If both the stronger and weaker mechanism would operate at its Cross-over Error Rate in parralel, the overall performance of the combined system would reduce. The use of the stronger biometric alone would be preferable in that case.
There is a huge difference in using two ore more biometrics in parallel or in series. In parallel both biometrics must be positive to conclude for a positive result and thus this is more difficult to accomplish.
When used in series, the first biometric can result in a number of possible candidates and the second biometric only checks the biometric against the enrolled templates of these possible candidates and thus resulting in more accuracy.
For some applications however, a very low false match (the customer service applications) or false non-match rates (the high security applications) is required, where a high opposite error rate is accepted. Combining biometrics in these cases could provide a very good result, however some issues are introduced: The main issues are:
· Combining biometrics could reduce the user acceptability, because enrolment is more lengthy and because it could increase the perception of invasion of privacy.
· The complexity of the verification process increases significantly; and
· The cost increases significantly.
Documentation on Multi Modal Biometrics, combining several forms of Biometrcs.
