Notices

I know there have been attempts in the past here to port OCR softwares like Tesseract to the N900. But all the threads that I have searched seems to have been abandoned for more than a year. So I am creating a new thread.

Right now, we have a lot of activity going on here development wise. Also, the Maemo/Meego coding competition is coming up. So please, could someone code a usable Business Card Reader? I use my N900 as my primary connectivity device and I sorely miss a BCR type app.

Hoping....

edit: my bad

OCR = Optical Character Recognition.

Would you mind to add a proposal to the Wiki of the Meego coding competition.

1. what is bcr?
2. tesseract is open source and available for linux, so why not use this?

edit: or better OCRopus

BCR- Business Card Reader
OCR- Optical Character Recognition

Yes tesseract and OCRopus are available for Linux. So it is just a matter of porting to maemo. I need help with that from an experience guru.
BTW, here somebody has already done substantial work, but no deb packages seems to be available.

Even some discussion seems to have happened here...

I am bumping this thread again, because I am now sorely missing this feature on the N900 and I deal with good amount of business cards in my line of work.
I found a link to a paper about text/image seperation in a BCR app:http://arxiv.org/pdf/1004.0766
It may help some one.

I am willing to donate for somebody to develop a BCR app for the N900. Also, you get a chance to participate in the Meego Competition and win something

and i'd also give this another bump.
hope somebody starts looking at it. and to that kind soul, thank you!

there is a package tesseract-ocr that you could try?

OK... so had a quick scooch on the debian package lists - it turned up a few things:
- tesseract. v3 (Sept 2010) is now on code.google The package on the debian site has dependencies that aren't armel compatible but that doesn't mean to say the code couldn't be compiled for armv7 though. Equally, it might not be possible.
- gocr. Looks less complete compared with tesseract but might be easier to get working with less depends.

It might be that many ocr programmes need to be run in a desktop environment so even if they can be compiled for armv7, it's only half the battle. Perhaps running it in ED might work?

There may be other foss projects that could be made to work on armv7...

Also, if TMO user jrocha reads this, please let us know how progress is on your OCR project.