Appeal No. 1996-2198
Application 08/077,506
The prior art Rabiner-Levinson speech recognition system
described at pages 2 and 3 of appellants' specification
("Rabiner-Levinson system"), employs an occurrence probability
distribution B that is a set of discrete probability values
obtained from quantized feature vectors. As is apparent from
the drawing identified as Appendix D to the brief (wherein the
reference numbers have the suffix "A"), the Rabiner-Levinson
system uses many of the same components as appellants'
invention, shown in appellants' Figure 1. Both systems have a
feature analyzer 10, 10A for analyzing an input speech pattern
to produce a time sequence of vectors V. When the invention
is in the training mode, the mode selection switch (13;
unnumbered in App. D) applies these vectors to a converting
2(...continued)
Accord, In re Raske, 28 USPQ2d 1304, 1304-05 (Bd. Pat. App. &
Int. 1993); Ex parte Hiyamizu, 10 USPQ2d 1393, 1394 (Bd. Pat.
App. & Int. 1988). See also MPEP § 706.02(j) (6th ed., rev 3,
July 1997)) ("Where a reference is relied on to support a
rejection, whether or not in a minor capacity, that reference
should be positively included in the statement of the
rejection. See In re Hoch, 428 F.2d 1341, 166 USPQ 406, [407]
n.3 (CCPA 1970).").
-5-
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Next
Last modified: November 3, 2007