Cuneiform OCR

[eslavko]

Hello...

Does someone know if there is commandline option for cuneiform?
I try to do OCR on my pages and tested tesseract and cuneiform. The other one have better result but lack of commandline to automate.

Slavko.

[dingodog]

*Tesseract* and *Cuneiform* for Linux are fully scriptable via command line

I built binaries of

*Tesseract 3.0* + additional language data in one file (213 MB)
- http://dokupuppylinux.tk/programs:ocr#tesseract-30

for Puppy Linux 3.01, 4.3.1, 5.2.5

and

*Cuneiform*
- http://dokupuppylinux.tk/programs:ocr#cuneiform_10
usage

Code: Select all

cuneiform [-l language -o result_file –html –dotmatrix –fax] < image_file >

for Puppy Linux 5.2.5

In the same page you can find also other ocr engines scriptable via command line

I'm currently working to build latest version (4.x) of OCROPUS, used by googlebooks, but it is a long and difficulty task, since dependencies are odd and soyrce code is not properly packaged, so I need to fix before to build

[eslavko]

I forget to say for WinXp...

For linux I know that is both command line options. (there is no GUI )
But for now I'm stuck in win. I run Ubuntu on other machine (EMC2 machine controller) but don't want to bloat it as is precise rtai environment.
So some solution for win?

I read somewhere that dll's can be acessed (puma or something similar) but doesn't find the parameters options.

[Misty]

Tesseract is available for Windows in official CLI form: http://code.google.com/p/tesseract-ocr/downloads/list There are also unofficial (probably unscriptable) GUIs, but I'm not familiar with those. Cuneiform for Windows is also available, but I don't know if it's the same CLI interface as the Linux/Mac version: http://en.openocr.org/download/