Gustavo Chain wrote:
> Hi Folks.
>
> I'm developing the Secugen-FDU2000 driver for fprint.
> Currently it show the finger image and make the match.
>
> I need testers for feedback (please)
Cool! How did you write the driver? (are specs available, did you
reverse engineer it, or what?)
Can you post the "lsusb -v" output for this device?
How well is imaging working? i.e. can you verify and identify
accurately? Roughly how many minutiae are detected per image? What's the
usual kind of bozorth3 match score for scanning the same finger twice?
I have not heard of this sensor before and don't think it's been
mentioned on the list, so I wouldn't be surprised if nobody has one. I
might be wrong though :)
> TODO:
> 1º Code cleanup.
> 2º Resample Image 398x300 -> 260x300
Why? Is this some kind of anti-distortion?
A few comments on the code:
> #include <errno.h>
> #include <string.h>
> #include <unistd.h>
>
> #include <usb.h>
>
> #include <fp_internal.h>
>
> #define FP_COMPONENT "fdu2000"
FP_COMPONENT needs to come before other includes in order to take effect.
> bytes = usb_bulk_write(dev, USB_ENDPOINT_TYPE_ISOCHRONOUS, cmd,
> sizeof(cmd), BULK_TIMEOUT);
Your use of the endpoint field is wrong. I really doubt this is an
isochronous endpoint, because you're using a function for a bulk
endpoint, and libusb doesn't support isochronous endpoints :)
Instead you should be using something like: 1 | USB_ENDPOINT_IN
> if (bytes < 0) {
> fp_err("%d: %s", __LINE__, usb_strerror());
> }
I'd prefer a meaningful error message rather than a line number, as
these will change between releases etc.
I'll leave further comments for when you've had a chance to do your
planned code cleanups. Thanks a lot for doing this, I look forward to
including it in libfprint :)
Daniel
_______________________________________________
fprint mailing list
[email protected]
http://lists.reactivated.net/mailman/listinfo/fprint
