Gerard Klaver wrote:
> Just did some test with fprint_demo version 0.04.
> Device Medion fingerprint AES2501 device
> 
> Question that i have is how much counted minutiae gives a correct
> fingerprint identify?
> Did some test but no matching yet.

Not quite sure what you're asking here. Minutiae are feature points 
detected on the fingerprint. It doesn't matter how many there are, as 
long as there aren't too few and there aren't an excess of invalid ones.

Are you saying you've yet to get it to correctly match a fingerprint? 
(i.e. always getting false rejections?)

Maybe you've got image quality problems similar to ones described here:
http://www.reactivated.net/fprint/bugs/task/2

> Wishlist:
> To have in the verify window a side by side view of the enrolled
> fingerprint and the verifyed fingerprint so an easier check can be done
> if the same manner of moving of finger is done?

Not possible at the moment, as enrolled fingerprints are not stored on 
disk, only the minutiae sets.
In theory we could implement the above using enroll-to-memory (rather 
than enroll-to-disk) and use the scanned image from the successful 
enroll stage, however this will not be possible in future as we will be 
switching a 3-scan enroll process.

At the moment you can use the "enroll" and "verify" libfprint example 
programs to do what you want. They both save images from every scan.

Thanks for your feedback.
Daniel
_______________________________________________
fprint mailing list
[email protected]
http://lists.reactivated.net/mailman/listinfo/fprint

Reply via email to