#MigraphOCR
3 messages in this thread
I have been using Migraph OCR, with pretty good results. On decent printed
or typed text, it can be significantly faster than re-typing. I am using
version 1.08. The first version I got did nothing but crash, so this is a
big improvement.
System: Amiga 3000 with 4 mb fast, 2 mb chip memory, A2286 Bridgeboard.
Problems I am still having:
Migraph OCR v. 1.08:
1. During "Finishing Phase" of OCR, it sometimes crashes. No error message
from program, just a guru 4 or 3 from the system. 2. During "Character
Recognition Phase" it sometimes crashes with a guru 5. 3. After first
starting the program and making a scan, the scanner usually turns on a
second time, so the first scan is lost. This only seems to happen on the
first scan. 4. There's no way to un-learn characters that it gets
consistently wrong, and is sure of. 5. I'd like to be able to create a
lexicon for a new language. Even if it were just a simple word list, it
would be helpful.
Migraph TouchUp v. 1.04:
1. I can't load an IFF file with more than 1 bitplane. TouchUp crashes with
a guru 4 or B. 2. I can't load a PrintMaster shape file. TouchUp crashes
with a guru 4. 3. I can't save a (properly sized) clip to a PrintMaster
file. It always says "Error while saving PrintMaster icons".
Are you leaving the OCR in the training mode? If so, it will continue to
add the same characters to its lex file until it runs out of ram without
warning. Once you've taught it a character, don't teach it the same
character again. And don't teach it on bad characters, just correct it.
Mike
Doesn't sound like "good" results to me! I just got the program, sent in
resistration card and hope for upgrade soon. I've just played a little with
it. Definately DOESN'T like none-letter dot-matrix. Also my Epson scanner
eats bytes; I have 8 meg fast, 2 meg Agnus and it can barely do 1/2 a page.
I guess the saying it true "you can't have too much memory".
– via Coolwhap