Our tests revealed that enrolment for this device requires planning and providing an adequate amount of buffer to the time planned. During enrolment, user rejection was rare. The results are good for an average company, but for larger ones, proper planning is unavoidable.
Although it depends on the actual usage, the time spent operating the device is a rather important parameter along with the ratio of (falsely) rejected identification attempts. Another question is how much time and effort is required for the enrolment of a user, and what percent of the possible users cannot be enrolled at all. Considering the large theoretical user capacity of the device, large amounts of data may be enrolled in a short time.
We sought answers for these questions and our experiences were mostly pleasant.
While the FNMR values can only be truly interpreted with other factors, our results show that the device has an order of magnitude larger FNMR rate than the factory disclosed value, almost quarter of the attempts. This is an average performance compared to similar devices.
Considering the ARAD values - the required time for a successful identification - the device performed very well, which allows for a fast and convenient operation for users. We found little difference between the ARAD values of successful and unsuccessful attempts.
Our stress tests revealed that the device do not cope with disturbing factors whether they result from environmental conditions or user behaviour. We do not suggest using the device in contaminated environments, as contaminating the sensor or the sample renders the device unable to successfully identify.
You can further read about the device in the following documents: