Appeal No. 1999-2294
Application No. 08/807,430
But, the examiner has not applied art to this concept, and we are under no obligation to
perform a comprehensive search of the prior art. Furthermore, with respect to claim 12,
we note that the Scott article (not applied) discusses the use of a two-pass analysis of
image data, not speech data, using both parametric and nonparametric models and the
benefits thereof in crop recognition. With respect to claim 1, Higgins teaches speaker
recognition using stored utterances of individual human speakers. We note that claim 1
does not even require speech recognition, but only assessing a degree of resemblance to
observations which may not require content recognition.
With this said, we agree with appellants that the prior art taught in Journal does not
2
teach the inventions as recited in independent claims 1, 12, 19 and 21 . Nor does the
combination with Scott remedy the deficiency in Journal alone. Appellants argue that the
invention is directed to using nonparametric speech models (individual/single utterances
by a single speaker) in speech recognition. (See brief at page 2.) Appellants argue that
Journal does not teach or suggest assessing a degree to which a speech sample
resembles a group of training observations (which are single utterances by a single
speaker) by evaluating the speech sample relative to particular training observations in the
group of training observations. (See answer at page 5.) We agree with appellants.
2We note that the examiner applies the combination of Journal and Scott to claim 21 with respect
to nonparametric models, but does not apply the same combination to claim 19 which contains similar
limitations.
5
Page: Previous 1 2 3 4 5 6 7 8 9 Next
Last modified: November 3, 2007