Rocr plugin

This Optical Character Recognition plugin enables Users to define their own databases of recognizable characters. It utilizes the Tesseract, Evaluate and ReFound programs.

An rocr database contains a 'Recognition Point' (RP) map for each character, built by the User. Most character maps contain just 4 or 5 RPs, making it quick, reliable and efficient method of recovering text from images. You are not even restricted to recognizing normal character images – you may also use it to assign a code value to glyphs of rotated text, symbols, logos etc. and then simply process the glyph as a regular character.

First create a sample PNG image of characters to be recognized, see Sampling.
Assign character codes to the glyphs in the sample image and set up their RP maps to create the database, see Training.
How to select RP pixels in the sample's glyphs to build an RP map: see Recognition Point maps.
How to process your document using the rocr database, export the recovered text as data-file and/or manipulate it using EscapeE and the evaluate plugin: see Process files.