Ocrocis

CIS projects: glp_pos, lowresCCWR, ne_bible, DC_DA, anchor-embeddings, lm-for-liic, unsupervised-graph-text, SherLIiC, MED, corefResources, noise-mitigation, complete list

About Ocrocis

Ocrocis is a project manager interface for Thomas Breuel's Ocropy, a toolkit for training optical character recognition (OCR) models with long short term memory networks.

Ocrocis guides the user in training high-quality OCR models from raw PDFs or images. It features automatic project management with a minimalistic philosophy:

Smart parameter inference, few required arguments
Non-redundant disk usage via hardlinks (Linux only)
Only core-Perl dependencies (apart from Ocropy's native dependencies)

Available for Linux and Mac OS X.

Details

Downloads

Ocrocis: https://github.com/kmnns/ocrocis
Tutorial: http://cistern.cis.lmu.de/ocrocis/tutorial.pdf
Tutorial data: http://cistern.cis.lmu.de/ocrocis/tutdemo.zip

If you have any questions, suggestions or bug reports, send an email to the developer: kaumanns [aet] cis.lmu.de

License

Ocrocis (2015) is licensed by David Kaumanns (Ocrocis software) & Uwe Springmann (Ocrocis tutorial) under Apache 2.0 (http://www.apache.org/licenses/LICENSE-2.0)