Hi, Are you trying to use identify or verify function? And the most important thing, are all your fingerprint sensors the same? If not, you can't use libfprint for it (or possibly any library)
you could possibly use pam_fprint and rewrite the part doing the fingerprint loading to load the data in any way you like, and also create some function to match the matched print to a user (if you plan to do complete identification) because the current form of pam_fprint loads all the prints for specified user and tries to identify against them. or even better, you should use fprintd, which is AFAIK the hot new stuff and thus not documented ps. sorry for my english and email format (i am new to this mail list thing) Pavel Walter Leibbrandt wrote: > Hi, > > As my first statement to the creators of fprint, I would like to > congratulate you on creating and maintaining this much needed library. > > I'm currently using libfprint to authenticate users in a system where > the data is stored in a central database. This seems to be an overlooked > use-case of libfprint, as I cannot seem to find a way (in the > documentation or examples) to validate fingerprints from data stored in > a database. > > The process of enrollment is only done once for a user and the > fingerprint (or minutiae) data is stored in a database. When a user > later tries to login at _any terminal_ running client software, the new > scanned data is used to validate against all fingerprint data in the > database. The enrollment process of libfprint is thus skipped and > handled by my application itself. I emphasize the fact that logins are > not necessarily done on the same computer where the user was enrolled. > > Ideally I would like to be able to scan a fingerprint from my > application and match it against the fingerprint data in my database. > This would require me to either have the scan and verify functions > separate, or to have a way for my program to enroll all data in the > database into its local gallery without having any actual scans performed. > > Is there any way for me to achieve this behavior, or would I have to > consider the major architectural changes needed to work with libfprint > as documented? > > > Thanks in regards, > > Walter > _______________________________________________ > fprint mailing list > [email protected] > http://lists.reactivated.net/mailman/listinfo/fprint > _______________________________________________ fprint mailing list [email protected] http://lists.reactivated.net/mailman/listinfo/fprint
