background image
C H A P T E R 9
About the Recognition System
the expansions of words that match entries in the expansion dictionary.
a graphical representation of the original input strokes as ink.
buttons for the soft keyboard and text-corrector views.
a Try Letters button when the cursive recognizer is active.
Figure 9-3
Text-corrector picker
The words in this list are one example of correction information stored by the
system as it recognizes words. In addition to word lists, correction information
includes the original stroke data and information known as learning data.
Learning data
is information gathered as the user corrects misrecognized words. It
is used to modify the system's internal handwriting model to more closely match
the way the user actually writes. This information is called learning data because
the system can be said to learn various characteristics of the user's handwriting
style, with a resulting increase in recognition accuracy. Not all recognizers return
learning data.
User Preferences for Recognition
The user can specify several preferences that affect the overall configuration of the
recognition system. This information is provided for reference purposes only;
generally, you should not change the user's recognition preferences settings.
© 2007-2017, o7 studio » при воспроизведении материала сайта ссылка обязательна