eugen at leitl.org
Mon May 13 16:21:53 UTC 2013
On Mon, May 13, 2013 at 10:08:57AM -0600, Kelly Anderson wrote:
> If this is something you deeply care about, I would recommend that you get
> both Omnipage and Paperport professional from www.nuance.com the industry
> leader in OCR.
> OCR is near and dear to my heart as it was the topic of my incomplete
> Master's Thesis. Nobody has yet developed the software I envisioned then.
> Too bad. I'm sure it will happen sometime though, but even if what I wanted
> to do wasn't implemented (it had to do with forms processing) what you want
> to do is more than adequately covered by the above products.
At my dayjob, among other things we machine-process millions of
documents (try e.g. Chemisches Zentralblatt, anno 1871). The
recognition rate is still not quantitative after combining
the input from three independant packages.
In case of chemical structure input, the task is Turing-complete.
Unfortunately, there is no information to flag failed input, in order
to have them processed by a human analyst.
More information about the extropy-chat