Appeal No. 1998-1386 Page 20
Application No. 08/269,156
Fujisaki does teach considering both character shape and
context information. Regarding the former, the reference
mentions, “a Frame Shape Recognizer, associates conditional
probabilities to each frame ....” Col. 13, ll. 67-68.
Regarding the latter, the reference states, “The language
model block 26 provides language model probabilities which may
be used to determine what characters are most likely to occur
in a given context, or what words are most likely to occur in
a given context.” Col. 9, ll. 23-27.
The examiner does not show, however, that the Frame Shape
Recognizer and the language model block operate substantially
simultaneously, let alone simultaneously. To the contrary,
Fujisaki suggests that the components operate sequentially.
Specifically, the reference characterizes the Frame Shape
Recognizer as one of the “pre-processing components ....”
Col. 13, ll. 63-66. Such a characterization implies that the
Frame Shape Recognizer operates before the processing of the
language model block.
Page: Previous 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Next
Last modified: November 3, 2007