[fprint] uru4000, number of enroll stages

[Martin Hejnfelt]

Den 25. jun. 2014, Vasily Khoruzhick <anarsoul at gmail.com> skrev:
> Hi Martin,
> > This fixes it perfectly! I can vouch for this change, at least it works
> > both with my U.are.u 4500 (using uru4000) and my upek touchstrip (using
> > upekts_img) readers... Thanks a lot!
> > One off-topic question: I got earlier reports that upekts_img doesn't work
> reliably on some devices. Does it work well for you?
> 
> Regards,
> Vasily
> Hi Vasily,

Well, first of all I guess it depends on what you mean by reliably.
I've not been using the upekts_img "in production" so to say. During the initial development of my current application I've been using it on my Lenovo X220 Upek touchstrip sensor, and an Upek TCS4K devkit, to test the fingerprint part of the software, where the sensor has worked fine, although matching accuracy was not great (which is also why I wrote a mail to this list some time ago, regarding which reader to use in multi-user environments). I will however not attribute the lack of accuracy to libfprint, as I've seen large installations with standalone strip based sensors where the accuracy was just as bad, thus I attribute it to strip sensors being bad for 1:n scenarios, due to it being very reliable on reproducible swiping. The U.are.U however to me, seems a bit easier as the user just has to press on the scan area, no movement involved.

For the record (and google) the current project involving the U.are.U 4500, is running on a Raspberry Pi with the git version of libfprint (added the patch Timo made ;) and matching seems quite good. I am a bit worried about speed when more users are being added, due to the RasPi's weak CPU, but I'll have to take it from there, or overclock the poor board :)

Regards,
Martin Hejnfelt

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/fprint/attachments/20140626/265f1209/attachment.html>