Optical Character Recognition (OCR)
Programs that read printed letters from a scanned page, first sold in Kurzweil's reading machine for the blind (1976), made paper text searchable and led to the mass scanning of books.
Open in the interactive tree →Reading aids existed from about 1914, when the Optophone turned printed letters into tones for blind readers, and in 1915 a device that converted special characters into telegraph code was patented. In 1976 Ray Kurzweil's company unveiled a machine for blind people that combined a flatbed scanner with a CCD (digital light) sensor, software that recognised nearly any typeface and a speech synthesizer; commercial OCR software followed in 1978 and the firm was sold to Xerox in 1980. OCR is how scanned pages become searchable text, including the over 40 million titles Google Books had scanned by 2019.
Prerequisites
- Stored-Program Computer (von Neumann)1945
- CCD Image Sensor1969The flatbed scanner of the first reading machine used a CCD sensor.
Unlocks
- Web Archives & Mass Digitization1996-2004Google Books turns scanned books into searchable text with OCR.
- AI Reads Archives & Inscriptions2022-2026Reading old documents by machine builds on character recognition.