Appeal No. 2006-2156
Application No. 09/790,296
We also sustain the Examiner’s 35 U.S.C. § 103(a) rejection of dependent
claims 2 and 10 which are directed to the feature of many-to-one mapping of the
phonetic vocabulary. We agree with the Examiner (Answer, page 8) that,
Appellants’ arguments (Supplemental Brief, page 8; Reply Brief, page 3) to the
contrary notwithstanding, the claim language does not require the many-to-one
mapping of a first language to a second language. As pointed out by the
Examiner, the language of claims 2 and 10 requires only that the mapping
applied to the phonetic vocabulary, which is built using the first language phone
set, be a many-to-one mapping which is described at column 2, section 2.3 of
Yuen. We also make the observation that Yuen’s Figure 4 illustration, which
depicts the mapping of Cantonese phonemes to English phonemes, shows that
multiple instances of Cantonese (first language) phonemes (e.g., “aai” and “ai”
and “au” and “aau”) are mapped, respectively, to the single English (second
language) phonemes “ay” and “aw.”
Turning to a consideration of the Examiner’s 35 U.S.C. § 103(a) rejection
of dependent claims 4 and 12 which recite a clustering feature for relabeling
feature vectors, we sustain this rejection as well. We find no persuasive
arguments from Appellants which convince us of any error in the Examiner’s
position as set forth at page 8 of the Answer. We agree with the Examiner that
section 2.5 of Yuen discloses the concatenating, i.e., clustering, of phoneme data
according to a first language phone set to obtain phonemes which do not exist in
the second language. Appellants attempt (Supplemental Brief, pages 8 and 9;
Reply Brief, pages 3 and 4) to draw a distinction between the claimed invention
6
Page: Previous 1 2 3 4 5 6 7 8 9 10 Next
Last modified: September 9, 2013