OCR Software

modmidget · #1 September 7th, 2007, 02:30 AM

I'm running Ubuntu 7.04 (Fiesty Fawn?) and I'd like to find a good OCR program. I've already looked at Tesseract but it can only be downloaded as a tarball and I've NEVER been successful at installing a program in that format. I'd like to find a good OCR program in .deb format that I can just double click on and watch the install happen. I tried installing Kooka but it requires the KDE desktop and would not run on my machine. I have Omnipage 11 loaded on my Win2000 computer but it doesn't work very well, so I'm hoping I can find something better for my Linux PC.

Does anyone have any suggestions?

**kage** · #2 September 7th, 2007, 07:01 AM

Checkout Clara, GOCR and Ocrad. Kooka also doesn't look bad, if you're using KDE (or don't mind using KDE/QT libraries).

modmidget · #3 September 7th, 2007, 05:18 PM

Thanks much for the links Kage. I checked them out and it looks like most of them are downloaded as tar.gz files, which never seem to work for me. I also did some additional research and found that the ocr process in Linux can be a real pain and very time consuming. The tesseract-ocr instructions I found on the web say that the user has to scan and save the document using sane, then you have to use something like Gimp to clean-up the file, then you have to do a command line instruction to convert the file to a .tiff format, then you can finally run tesseract from the command line to convert the file to a txt format. By the time your finished you have 3 files for a one page document, you end up with a bmp/jpg file, a tiff file, and a txt file.

I was hoping to find something that worked more like Omnipage where you just put your document in the scanner, click on aquire, and then click on the OCR button. I don't do a lot of OCR stuff so I think I'll just stick with Windows Omnipage.

**kage** · #4 September 8th, 2007, 12:35 AM

You're using Ubuntu. Don't download the source packages (tar.gz / .tar.bz2) of those programs, just use apt-get or synaptic (gui frontend for apt-get) to install them. Installation shouldn't take you 5 minutes.

modmidget · #5 September 8th, 2007, 02:43 AM

One hour ago I didn't even know what synaptic was/is...... it only took a few minutes to figure it out and less than 5 minutes to install Tesseract. Thanks once again Kage. Now I'll have to figure out how to use Xsane so I can get it to scan a whole document. I've scanned a one page document 3 times and I only get the left half, not the entire page.

If you have any suggestions on how to fix that problem too Kage they would be appreciated.