If i got the meaning of that parameter, you are lowering the number of minutiae needed to have a match to 10; indeed fingerprint verification succeeds most of the times now!
How did you choose "10" as a value? Is it secure, in a way it doesnt produce false positives? For instance here in italy the local government has set that value to 16-17 minutiae (with 17 being a 100% positive match) Thanks a lot for your help! guido 2009/7/31 Michael Cronenworth <[email protected]> > On 07/27/2009 06:06 AM, Guido Grazioli wrote: > >> Hello all, i'm new to this list; >> >> I have a somewhat working aes1610 driver for libfprint-0.1.0-pre2, >> based on the aes2501 driver and aes1610 driver from 0.0.6 >> >> I need some help in testing, as even if acquired fingerprint scans >> look good in fprint_demo, i cannot get any matches >> The driver also randomly crashes when swiping the finger >> too slowly in the reader. >> >> > Good news everyone. > > I found the missing parameter that allows matching with the AES1610 - > borrowed from the old aes1610 driver. I enrolled my right index and verified > it successfully several times. > > You'll find the patch against your source attached. > > The crash is in nbis/mindtct/dft.c:188 and I couldn't trace down what was > causing it but it may be a bug in libfprint and not the driver. I was not > able to reproduce it after applying my patch though. > > Thanks for pursuing this. Looks like my original patch was futile. :) > > -- Guido Grazioli <[email protected]> Via Parri 11 48011 - Alfonsine (RA) Mobile: +39 347 1017202 (10-18) Key FP = 7040 F398 0DED A737 7337 DAE1 12DC A698 5E81 2278 Linked in: http://www.linkedin.com/in/guidograzioli
_______________________________________________ fprint mailing list [email protected] http://lists.reactivated.net/mailman/listinfo/fprint
