[fprint] upek 147e:2016 reverse engineering progress

[Pavel Herrmann]

Just a thought, is the only difference between UPEK TS scanner variants 
the existence of the crypto coprocessor? Can we try to bypass the 
coprocessor and turn the "old" variant supported by upekts driver into 
imaging device as in windows with the advantage of identification?

I would definitely try but I don't have the hardware and don't know 
anyone with it using linux

Pavel

Daniel Drake wrote:
>>> I just spent some time sniffing what UPEK's new libbsapi is doing, 
>>> and produced a libusb-1.0 app to do the same. It works - it captures 
>>> image data. I'm not totally clear on the format, but the output file 
>>> looks quite nice when plotted as 8 bit greyscale with a width of 
>>> 2306 pixels (although the fingerprint is repeated).
>>
>> Updated version attached which works from cold boot. Also it now uses 
>> the hardware's finger detection capability and deinitializes the 
>> device after the scan.
>
> Another version which includes finger-removal detection and outputs 
> coherent images in pgm format. yay.
>
> That was less effort than expected. I've mostly finished converting 
> this to a libfprint driver as well, expect that in the next few days.
>
> Daniel
> ------------------------------------------------------------------------
>
> _______________________________________________
> fprint mailing list
> fprint at reactivated.net
> http://lists.reactivated.net/mailman/listinfo/fprint