Appeal No. 2006-2156
Application No. 09/790,296
and Yuen by asserting that the claimed invention is concerned with grouping
alignments while Yuen is merely refining mapping. We do not find this
persuasive since we find nothing in the claim language which precludes mapping
refinement. We agree with the Examiner that, while Yuen does indeed teach
mapping refinement, such mapping refinement accomplishes the realignment of
speech phonemes by concatenating (clustering) previously mapped phoneme
data (Yuen, section 2.5).
We further sustain the Examiner’s 35 U.S.C. § 103(a) rejection of
dependent claims 5 and 13 based on Yuen. While Appellants contend
(Supplemental Brief, page 9; Reply Brief, page 4) that Yuen fails to disclose the
separating out of labeled feature vectors for different phones in a first language,
we agree with the Examiner that the language of claims 5 and 13 does not
require feature vector “separation” but, rather sets forth feature vector
“clustering”. As asserted by the Examiner (Answer, page 9), we find that Yuen,
at section 2.5, discloses the sequential comparison of phonetic spellings of
lexemes associated with a first language phone set with lexemes of a grouped
second language phone set.
We also find to be unpersuasive Appellants’ arguments with regard to the
Examiner’s obviousness rejection of dependent claims 6, 14, and 18 in which the
admitted prior art is added to Yuen to address the “large vocabulary” feature of
these claims. We find no error in the Examiner’s assertion of obviousness to the
ordinarily skilled artisan of applying the teachings of Yuen to large vocabulary
continuous speech recognizer systems in view of the admitted prior art for the
7
Page: Previous 1 2 3 4 5 6 7 8 9 10 Next
Last modified: September 9, 2013