ocr?

Category: Geeks r Us

Post 1 by b3n (I'm going for the prolific poster awards!) on Thursday, 24-Jan-2008 15:38:49

Hi.

After I rescued a scanner from a bunch of computer parts that were destined for a tip, i've become interested in ocr and what it can do for me, but sadly, my knolidge is some what lacking atm.

At the moment, i'm playing with a frontend to an ocr engine which is open source and maintained by google and am getting mixed results with it.

I've heard from people that the products from abby are good and this is what the resources department use at school I think.
As far as programs go, has anyone used abby finereader? Obveously i'm concerned about ease of use, but at the end of the day, if one program uses a better engine than another that is easier to use, i'll go with the better engine.
I'm also not overly sure about scanners and what settings to tweek.
The scanner that I found is an epson perfection 1200, which google suggests is about 8 years old, but i'm still getting readable output for the most part from it.
Atm, when i'm scanning stuff, the program defaults to 150 dpi, turning this up to 1200 which is the most that the scanner supports seems to have no effect other than longer scanning times - I thought the higher the dpi the better the image?
I can also change things like contrast, but i dont have a clue what values to set this to, so i've left it alone.

So if someone could fill me in on any part of the above i'd be greatfull.
Chears, BEN.

Post 2 by louiano (I'm going for the prolific poster awards!) on Friday, 25-Jan-2008 3:46:58

I'd use openBook instead. Unfrotunately its not that afordable... i am not sure what I'd use if I didn't have it. If its a pdf you can actually print to text using the generic text printer driver by installing or adding a new printer. Your output will be a .txt file on the folder you choose.

Post 3 by wildebrew (We promised the world we'd tame it, what were we hoping for?) on Friday, 25-Jan-2008 6:00:39

You should be getting pretty good results with printed English texts, I've used a flatbed scanner with Omnipage and I'm getting pretty good results, it's an HP scanner I think (just use it at work sometimes).
Omnipage used to be the best but it's kind of annoying now for OCR. Scanners are cheap though, sub $100 range usually, and they often come with passable OCR software that can also read pdfs, there's a Polish brand of OCR software that si supposed to be good, may be it's abby, can't remember.

Post 4 by blbobby (Ooo you're gona like this!) on Friday, 25-Jan-2008 6:09:24

Haven't used abby in years, but when I did you had to use the jaws cursor a lot (unconventional dialogues), but, if you are patient it can work.

Post 5 by Eleni21 (I have proven to myself and the world that I need mental help) on Friday, 25-Jan-2008 11:33:54

I've been using scanning software for about six years now. I've tried Kurzweil 1000, Cicero and Open Book. I'm mostly used to Kurzweil, but bought Cicero because it could handle Greek text. However, it can't scan two languages on the same page. I've yet to find software that can. As for ease of use, I'd still recommend Kurzweil. That's a rarety for me, as I usually side with Dolphin on most things, but I find that the settings in Kurzweil make more sense. I haven't worked with Open Book in quite awhile but they seemed straightforward as well. Btw, Finereader drove me nuts and I wouldn't recommend it to anyone who needs a screenreader.

Post 6 by b3n (I'm going for the prolific poster awards!) on Saturday, 26-Jan-2008 21:44:23

thanks for all the responces.
Anyone kno about the dpi thing?