Does anybody know the FineReader OCR program? I need some tips and tricks to improve the recognition rate.
For a job, I need to read ca. 100 pages of multi-language, two column text in text script, italics and boldface with this program. I've done ca. 25% already, with error rates of ca. 5 to 9 %. Most of the languages are written in Latin script, with lots of diacritical signs, especially macron, diaresis above and below characters, accent d'aigue, caret, hacek, single dots above and below the characters, but there are also some words, phrases, whatever in Chinese (traditional) and Arabic script. If I add the last two to the list of the current languages, the error rate increases, so, if I have Chinese signs or Arabic words, I make two runs, one with that languages for recogniton, the other without them for reading the rest. I've also already used the teaching feature. I made the observation that reading it a few times will give better results that just using the first run.
Discussions worth keeping around later.
1 post • Page 1 of 1