ocre : o.c.r. easy (and free/libre)

v0.039

download

ocre...tgz (on linux).
debian: in /etc/apt/sources.list you can add:
deb ftp://lem.eui.upm.es/pub/lemdeb squeeze main
deb-src ftp://lem.eui.upm.es/pub/lemdeb squeeze main
ocre...deb (id. i386, Debian).
ocre...rpm (id, RedHat, ..).

ocre works with grey images

Help

Languages:

English, Euskara/Basque, French, German, Polish, Português, Russian, Spanish

License:

GPL

input: PGM/PBM file
output: unicode or ASCII characters in standard output
process:

segmentation, and more
comparison
if new: take measurings and consult knowledge base
if unknown: ask user

This version:

(39) First tags in the output, page number. (ocre -y1 ...).

Next versions:

some errors corrected, better recognition. :-)
esperanto, more dictionary, multiple columns, books opened, more cyrillic, simple zoning,

You can subscirbe to new releases.

lacks (wants)

revision (better use of dictionary, trigrams,...)
robustness
table recognition, formula id., spreadshet id., ...
zoning
interface with:

gnome, xml
scanners
...

output to:

latex, html, gnumeric, ...

...

In the long term

I want to teach the computer to read. :-)

Webs OCR

UC Berkeley CS 294-9 - Document Image Analysis
University of Nevada, Las Vegas Information Science Research Institute
University at Buffalo Center of Excellence for Document Analisys and Recognition.
OCRchie Modular Optical Character Recognition Software
OCR resource list

Acknowledgements

... It's a long way to go

Updated by Luis José Cearra Zabala
jue jul 28 21:02:48 CEST 2011