Quote:
Originally Posted by kodomo
I'm personally quizzy as to whether one wants to do OCR on an embedded ARM device with 256MB RAM... but ok... :P  |
Of course I don't mean a full-fledged OCR system, ARM is indeed too slow for that. But for the purpose mentioned by Innit it would be useful. Imagine you have a scanned pdf and you want to highlight a fragment of text and save it, just as you do with regular pdfs. With OCR option as I see it, you could select a rectangular fragment of the scanned page image, and OCR just it, not the whole page. This is very useful particularly for bibliography retrieval.
I used exactly such scenario on a slow Intel Atom netbook, with a silly little bash script (Tesseract's output piped to Google Scholar to retrieve a Bibtex reference), and it worked flawlessly. So it should work on Boox too.
Quote:
Originally Posted by kodomo
Go ahead - the Wiki's there (and by no means mine alone )
Cheers! |
I'll consider that.