Re: Finding Characters above "z"
- Hi Alec (,Hugo, Dan)
--- In notetab@y..., "Alec Burgess" <burale@a...> wrote:
> Your spoiling the fun of an terse and elegant but useless
> Regexp expression :-(
> The only OCR I've used much is an old (but powerful) version
> of Textbridge and that mostly the odd time for converting
> screenshots to text. My experience with Textbridge was that
> it had a "training" mode and that after doing the first
> couple of paragraphs almost one character or word at a time
> it would be able to complete the remainder of the job almost
> letter perfect.
> My point being that if your OCR makes mistakes consistently
> its probably possible to train it to do it right. Is this
> not so in the "real" world?
Yes, if it can be re-recognized without rescanning.
But come to think of it, I've had my best result cleaning up
OCR errors with a good spell-checker; when it finds a
misspelled word you can designate the correction (often by picking
it from a list) and "Change All."
The spell checker in Notetab Pro might be good choice. It seems to
have those features. (My experience was with MS Word tho).
> ---- Original Message ----
> From: <h.paulissen@f...>
> To: <email@example.com>
> Sent: 20 September, 2002 04:02
> Subject: RE: [NTB] Finding Characters above "z"
> > Hi Dan (,Sheri, Alec),
> > A lot has been sain about regular expressions for doing
> > this, but I doubt whether it is the best solution.