Appeal No. 95-1776
Application No. 08/096,686
Reference is made to the brief and the answer for the
respective positions of the appellants and the examiner.
OPINION
In response to the examiner’s obviousness determination
(Answer, pages 2 through 5), appellants argue that Baker does not
disclose ‘means for generating and storing . . . a data record’
because “the vocabulary language model and word models are data
structures which are stored prior to providing the speech
recognition functionality” (Brief, page 6), and that Baker does
not disclose a “historical database based on the successive
recognition events in a dictation session” (Brief, page 6).
To the extent that multiple candidates are identified in Baker
(Figures 26 and 27) as part of the recognition process,
appellants argue that “they are used only in permitting
correction of the current word and there is no suggestion of the
creation of a database wherein each record identifies the best or
most likely candidate and alternate candidates for each
recognition event” (Brief, page 6).
In Baker, the multiple candidates in display 478 (Figure 26)
are stored in memory before a “spoken input event,” and the only
“data record” that is created in Baker (column 18, lines 53
through 56) as a result of the “spoken input event” is the best
selected word (i.e., DEMONSTRATES) output 482. No other data is
stored in Baker to generate a historical “database” for
Page: Previous 1 2 3 4 5 6 7 Next
Last modified: November 3, 2007