Pocr

Pocr is a program for preprocessing images. It is meant to be run on the digitised pages before the OCR, and is targeted a workflow with a BookDrive DIY camera based scanner.

Features:

• Automatic rotation of odd/even pages
• Perspective correction, with interactive transform and propagation of changes to the following pages
• Creation of gimp-script for generating both OCR-ready and web-ready images

Some of the transformation code is released under GPL, as it is based on similar code in the Gnu Image Manipulation Program. The remaining part of the code is public domain.

Status

The software is still incomplete and not ready for production use.

The project page is located at http://sf.net/projects/pocr/.

Screenshots

Credits

The software is developed for a digitisation project funded by Biblioteksstyrelsens Udviklingspulje for Folke- og Skolebiblioteker, Kraks Fond and Københavns Hovedbibliotek.

Perspective tranformation code is based on code from the GNU Image Manipulation Program. Fltk 1.1 is used for image loading and graphical user interface. Valgrind is used for memory profiling.

Hosted at: