Reading images

Abstract

Optical Character Recognition (OCR) is the digitisation of printed pages into editable and searchable computer readable texts. OCR software analyses patterns in an electronic image of the page to work out which letters and words are in the document. A major issue is quality. Even 99% accuracy would give 2 errors every 3 lines. One form of quality control is proof reading, but this is expensive. An alternative is Adobe Capture which replaces suspect reads with a ‘bitmap’, a picture of the original.

Citation

Pitcher, G. (2000), "Reading images", VINE, Vol. 30 No. 1, pp. 39-41. https://doi.org/10.1108/eb040739