Appeal No. 1998-1902
Application No. 08/435,592
improves upon the Floyd and Steinberg filter (Fig. 5A) by changing the rightmost weights
(7/16 and 1/16) to 8/16 and zero (Fig. 5B). Thus, in the terminology of instant claim 1, the
Floyd and Steinberg filter has weights of 7 - 1 - 5 - 3 expressed in clockwise sequence
from the origin (i.e., the black dot which represents the present pixel being scanned).
Barton’s filter may be expressed as having weights of 8 - 0 - 5 - 3 in clockwise sequence
from the origin. Barton also makes reference to error diffusion filters having more terms
than the four contained in the Floyd and Steinberg filter (see column 6, lines 13-19.)
However, we do not find any suggestion in Barton for the filter having weights of “8 -
0 - 4 - 4 expressed in clockwise sequence from the origin,” as required by instant claim 1.
Nor has the examiner presented any convincing rationale that shows the subject matter as
a whole of claim 1 would have been prima facie obvious to the artisan.
We acknowledge that Barton, with reference to Figures 5A and 5B, teaches moving
weight from the 1/16 (lower right) box to the 7/16 (upper right) box, such that the upper right
box contains weight of 8/16, with the lower right box containing weight of zero. In
retrospect, we also see that moving 1/16 of the weight in the 5/16 (lower center) box to the
3/16 (lower left) box of Figure 5B would result in the filter claimed -- 8 - 0 - 4 - 4. However,
absent this impermissible hindsight, we fail to see, based on the evidence before us, how
the artisan would have been led to select the claimed parameters.
Appellants’ specification provides motivation for the departure from the prior art.
As stated in the “Summary of the Invention,” having “one of the weighted error terms twice
-4-
Page: Previous 1 2 3 4 5 6 7 8 9 Next
Last modified: November 3, 2007