CJKV Information Processing, 2nd Edition

354

|

Chapter 5: Input Methods

physically wrote what was intended. Pen input depends on another technology, speci-

cally OCR, to be covered in the next section. Perhaps of historical interest, GO Corpora-

tion had enhanced their pen-based OS to handle Japanese through the use of Unicode.

Microso’s MS-IME, which was the standard input method provided with various

versions of their Windows OS, provided the user with the ability to input characters by

writing them on a special onscreen tablet.

Optical Character Recognition

Several OCR systems currently accept CJKV character input, although there are, of course,

limitations. e clearer and larger the typefaces, the more reliable such a system is. Some

systems do not recognize all the characters in a character set (in the case of GB 2312-80,

for example, some recognize only Level 1 hanzi), and some are restricted to certain type-

face styles.

You encounter OCR systems more frequently in the West where recognition of a much

smaller collection of characters is done. e recognition of thousands of individual char-

acters becomes much more dicult, particularly when each one is fairly complex in

structure.

NeocorTech’s KanjiScan OCR, available only for Windows, can convert printed material

that contains Japanese and English text into a form that can be manipulated as normal

text.

*

For those characters that it cannot recognize, ...