Open Source OCR, an ABBY alternative?

[daniel_reetz]

Background: http://www.teleread.org/2008/05/26/amaz ... e-sourced/

Homepage:
http://en.openocr.org/

Forum thread with built packages:
http://openocr.org/forum/viewtopic.php? ... 63&start=0

[Kirtai]

Since it has hocr output, http://jimgarrison.org/moz-hocr-edit/ might be useful too

[monday2000]

It would be nice to adopt CuneiForm to OCR the DjVu files. Exactly to say - to convert the OCR results to a format of the DjVu-TXT layer.

Full details here:

http://openocr.org/forum/viewtopic.php?f=7&p=4417

Are there any volunteers?

[daniel_reetz]

Are there any volunteers?

You know our policy here, it's stated in our URL. Do It Yourself!

[monday2000]

daniel_reetz

Do It Yourself

I prefer to focus on such things that wouldn't be done by anyone else (as for DjVu-scanning etc.) - that's more efficient than a plain "Do It Yourself" strategy.

[qwer]

This http://www.onlineocr.net/ service uses tesseract http://code.google.com/p/tesseract-ocr/.
Training process described here: http://code.google.com/p/tesseract-ocr/ ... gTesseract .

[Tim]

Some one already mentioned tesseract http://code.google.com/p/tesseract-ocr/, its currently said to be the most accurate open source OCR, but it only does one column text. You have to do the layout analysis separately to break the source document up into portions that are only single column text. Tesseract is being developed, but not rapidly.

The most rapidly developing open source OCR system is ocropus. http://code.google.com/p/ocropus/
It has some of the leading image processing research people working on it. If you follow the names of the people that release papers on binarization, ocr, etc, the people working on ocropus are up there. It's alpha code right now, but it does layout analysis and can optionally use tesseract as a plug in.

I'm actually building cuneiform right now. It does appear to have layout analysis, so it may be farther ahead than I had heard from other information. It looks promising.

But in short none of the open source systems are as accurate as ABBY Fine reader or Omnipage, but they may get there.

[monday2000]

But in short none of the open source systems are as accurate as ABBY Fine reader or Omnipage,

Yes, sadly.

but they may get there

Only if Web-communities are going to contribute. Hard, but not desperate.

[gabossy]

What are the advantages from a software developer's standpoint of making open source software? I am new to the tech industry (I am in the legal department) and am wondering from a practical standpoint why a software developer would develop open source software. Any thoughts?
________________________
keyword research ~ keyword tool ~ keyword tracking ~ affiliate elite

[Tim]

What are the advantages from a software developer's standpoint of making open source software? I am new to the tech industry (I am in the legal department) and am wondering from a practical standpoint why a software developer would develop open source software. Any thoughts?

That's a really big question, not necessarily for this forum, but from the the OSI's website:

Open source is a development method for software that harnesses the power of distributed peer review and transparency of process. The promise of open source is better quality, higher reliability, more flexibility, lower cost, and an end to predatory vendor lock-in.

You can also look at what the Free Software Foundation has to say about it, but they aren't very good about explaining the practical why. Wikipedia's article has some more of the why's but it's not all that well written.