3.13. Character Recognition.   

The other end of graphics is character recognition : given a graphic shape, recognise the character it represents (and cater for different fonts, typing mistakes,etc). This requires sophisticated programs and a means of graphics input ( typically a scanner or camera). The text or images can then be manipulated by the computer (see Figure 3.16).

Ray Kurzweil created the Kurzweil scanning system which can "read" and digitise typewritten or type-set text such as documents, 5 to 20 times faster than a keyboardist. It is an omni-font optical character recognition system (OCR) with programs developed using artificial intelligence techniques: it is capable of "learning" a typestyle by analysing the character shapes.

It scans enough text to develop standard shapes for characters including bold, italicised, etc (see Figure 3.17) and can then process the rest of the book or document only needing human intervention to turn pages or if it is unable to recognise a character to help determine what it is