in order to increase the chances of getting blurred pictures right in terms of currency (often misinterpreted),

can you please "teach" the OCR engine to (primarily) use the existing currencies list (as defined, or already used)

for instance, I cleaned my currencies dictionary of those I do not (expect to) use or
even better, increase the likelihood of the ones already used before on existing expenses, based on hit count, values, whatever can run in the background and not slow the process/ increase the load

so that, when in doubt, the system defaults to frequently used currencies (or even the base one) rather than relying on exotic ones.
I just wish I belonged to your top x% globetrotters :)
Thanks,

Tibi