SYNOPSIS

ocroscript <script> <arguments>

DESCRIPTION

You can see a list of all available commands by looking in the $OCROSCRIPTS (/usr/share/ocropus/scripts/ by default) path.

The 'recognize' script uses tesseract for recognition and sends the html-based hOCR ouput to stdout. Tesseract is probably the most mature text recognizer within OCRopus at the moment. Natively, Tesseract doesn't do layout analysis, but combined with OCRopus, it makes for a pretty good OCR system:

$ ocroscript recognize page.png > page.html

Here is a brief summary of the remaining command line commands available. You will need to look at the script to see what the command line arguments are:

degrade.lua

Simple document image degradation

hocr-to-text.lua

Convert hOCR output to plain text.

line-clean.lua

Given a line image, remove marginal noise and fix some other problems.

sauvola.lua

Perform Sauvola thresholding.

AUTHOR

ocroscript was written by Thomas Breuel.

This manual page was written by Jeffrey Ratcliffe <[email protected]>, for the Debian project (but may be used by others).