Appeal No. 1998-1902
Application No. 08/435,592
the magnitude of the other weighted error term, the present technique provides certain
computational advantages.” (Specification, page 8, lines 11-13.) “Each of the two weights
for the error terms is a multiple of 2 so that a shift in a binary shift register can be utilized to
obtain each of the weighted terms.” (Id. at lines 14-16.)
As for the former stated advantage, Barton’s reduction of processing time “by
nearly 25 percent” appears to refer to skipping one division operation with respect to each
pixel, since one of the weighted terms becomes zero, or null. Barton does not disclose a
“computational advantage” in having one of the weighted terms twice the magnitude of the
other weighted term.
As for the latter stated advantage, Barton does not disclose details of the circuitry
for performing the filter operations. The filter circuitry would appear to be substantially no
different from that of the prior art, represented in Barton’s Figure 4 simply as the “Error
Filter.” There was thus no recognition in the reference, and no suggestion, that using terms
that are a power of two would result in faster and simpler operations, allowing use of
simple binary shift registers to perform the required division, as shown in appellants’
Figure 1.
The vaguely-referenced “well known prior art” does not allege any particular
knowledge within the understanding of the artisan, and does not serve to lead to any
motivation for arriving at the subject matter of instant claim 1. While we agree that the
artisan would have been expected to seek to improve upon, and likely change, the filter
-5-
Page: Previous 1 2 3 4 5 6 7 8 9 Next
Last modified: November 3, 2007