OCR

7/7/2001

Optical Character Recognition. A technology that examines images of text and converts them from pictures to ASCII or other formats suitable for use with a Word Processor or other text oriented computer application. Intelligent character recognition (ICR) is an extension of OCR that extends OCR to handwritten character recognition.

300 dot per inch scanning with 256 gray levels is adequate except when dealing with very small typefaces. It is sometimes possible to do OCR with lower resolution and simple back and white. Flatbed scanners generally produce better results than handheld scanners.

OCR results can vary from nearly perfect to pretty awful. Fancy typefaces, multi-column text, boxed off text, etc. can contribute significantly to errors in character recognition. Spacing is often altered in tabulated data.

Return To Index Copyright 1994-2002 by Donald Kenney.