i've made quick port of tesseract:

http://brain.umcs.lublin.pl/~rzookol/download/tesseract_beta.lha

syntax:
tesseract input.tif output.txt <-l language>

before use download Your language data and coopy it to data/tessdata

http://code.google.com/p/tesseract-ocr/

This is great.

Let me guess: It will be integrated in scandal?

;)

[ Edited by eliot on 2009/9/19 18:12 ]

How do we convert a scanned PDF? In these files, each page is a jpg, i believe.
I suppose that we extract the jpg pages by using the PDFTools and then we batch convert each jpgs to tifs? Which tool is recommended for the batch convertion?
Is there a better method from what i proposed?

i planned to release more polished version (with jpeg/png) support but i have small problems with my vertebral column and stayed all day in bed :) . when i will be able to sit on the chair i'll finish port. I thinking about making sketch plugin (and add plugin api to scandal) or make standalone application.

btw. Upcoming tesseract version 3.0 (not yet available) will have png/jpeg and layout support.

@AmigaMancer
i've not tested convertiing pdf. I've tested scans from scandal only (150 dpi or more gives nice results)

Get well soon, Michal. Great job!

@ zukow

I wish you all the best, get well soon!

About the plug-in for Sketch, Scandal, perhaps others, this is a great idea, please do it!

Regards

Christoph