Training data for Document Vision


Training data for DocumentVision, a node.js library for processing and understanding scanned documents.

The training data is packaged this way as it very seldomly changes. The set includes:

  • English language support ('eng')
  • German language support ('deu')
  • Data for OSD mode

The module itself only exports two strings: filename contains module.filename and tessdata points to the tessdata directory (fully resolved).

DocumentVision is maintained under the Semantic Versioning guidelines as much as possible:

  • Version number format is <major>.<minor>.<patch>
  • Breaking backward compatibility bumps the major (resetting minor and patch)
  • New additions without breaking backward compatibility bumps the minor (resetting patch)
  • Bug fixes and other changes bumps the patch

Licensed under the incredibly permissive MIT License. Copyright © 2012 Christoph Schulz.