[fprint] Experiences with libfprint in multi user environments

[Mariusz Ciszewski]

Hello Martin.

I'm also trying to find working sollution you are talking about.

Unfortunatelly I am not able to write it by myself.
I'm also user / owner of a Raspberry Pi ( http://rasp485berry.wordpress.com/)

Only one I found it was a package for verification of a known user
(fprintd-enroll connected together with fprintd-verify).
But as we know It's not usable for identification.

I'm using: ID 08ff:2580 AuthenTec, Inc. AES2501 Fingerprint Sensor

My experiences on Raspberry Pi with fprintd 0.5.0-1 are very well (was very
bad under version: 0.4.1-5)
http://lists.freedesktop.org/archives/fprint/2014-March/000561.html

No false errors under 0.5.0-1 (fprintd-enroll connected together with
fprintd-verify)

If you have some software I can test it under my Raspberry Pi and AES2501
Fingerprint Sensor

Best regards
Mariusz Ciszewski
6 kwi 2014 08:13 "Martin Hejnfelt" <martin at duffman.dk> napisał(a):

> Hi Mariusz,
>
> I am writing my own program using libfprint and enroll/identification
> works with the touchstrip sensor I've been using, however I feel I have too
> many false negatives, and that will upset the users, which is why I want to
> know if anyone had experience with more expensive devices like the ones I
> mentioned in my previous post. The touchstrip sensor seems to be best in
> 1:1 scenarios and I need something that is good for 1:n.
>
> Kind regards,
> Martin Hejnfelt
>
>
> -------- Oprindelig meddelelse --------
> Fra: Mariusz Ciszewski
> Dato:06/04/2014 01.49 (GMT+01:00)
> Til: Martin Hejnfelt
> Cc: fprint at lists.freedesktop.org
> Emne: Re: [fprint] Experiences with libfprint in multi user environments
>
> > I am currently working on a unit where I need fingerprint
> > identification of up to 50 users
>
> Hello Martin
>
> Please let us know how it will work. You wrote you need identification, so
> verification (ready to use fprintd-enroll and then fprintd-verify) will not
> help you / will not enoug for you.
>
> I'm also searching for sollution where only fingerprint is needed (no
> keyboard needed, no user name berofe scanning is needed. I think it could
> be called pfrintd-identify.
>
> Please let me know if you will have some results.
>
> Best regards
> Mariusz Ciszewski
>
>
> 2014-04-05 14:15 GMT+02:00 Martin Hejnfelt <martin at duffman.dk>:
>
>> Hi,
>>
>> I am currently working on a unit where I need fingerprint identification
>> of up to 50 users. The system is going to run on Linux (right now with a
>> Raspberry Pi as host) which is why I use libfprint. Installing/compiling
>> newest libfprint (from the git repo) has been done.
>>
>> I have good experience with the UPEK touchstrip sensors, however in
>> multiuser environments I have the feeling that, first of all the number
>> of false negatives are a bit high, probably mostly due to how small the
>> sensor is, thus increasing the risk for people to make an "erroneous"
>> swipe. I feel that for an application to use fingerprint, for people to
>> not get "pissed" at the system, it must be rather stable.
>>
>> Therefore I am calling for some experience regarding the different
>> readers supported by libfprint regarding to stability and ease of use
>> (for the user), before I just start randomly buying readers :)
>> My best guesses right now are the URU4500 and EikonTouch 300 (are the
>> 500/510's supported?) since they use "touch" type which I guess would be
>> easier for the user, and give more minutiae easier, but this is more a
>> guess, since I only have experience with the touchstrip versions.
>>
>> Any help is appreciated!
>>
>> Kind regards,
>> Martin Hejnfelt
>>
>> _______________________________________________
>> fprint mailing list
>> fprint at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/fprint
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/fprint/attachments/20140406/b70ecd72/attachment.html>