Appeal No. 2006-1618
Application No. 10/046,797
when the scale factor λ=1 [Makram-Ebeid, col. 1, lines 60-62]. The scale factor λ
is then increased to 2, and the process repeated. Ultimately, the goal of
Makram-Ebeid is to merge the regions using the lowest value of the scale
parameter λ [Makram-Ebeid, col. 2, lines 8-10]. Thus, as the value of the scale
parameter increases, the regions are merged until the Energy function cannot be
minimized further [Makram-Ebeid, col. 2, lines 13-15].
In our view, Makram-Ebeid's teaching of eliminating contours is
reasonably combinable with the teachings of Catros essentially for the reasons
stated by the examiner. Furthermore, we agree with the examiner that Catros'
use of gradient amplitudes in an algorithm to determine the most appropriate
path to connect a disjointed contour reasonably constitutes numerical values
associated with a gradient and therefore weight values.
Furthermore, Makram-Ebeid's goal of merging regions by using the lowest
value of the scale parameter λ reasonably suggests selecting an optimal scale
parameter from a plurality of scale parameters by determining a scale parameter
that minimizes variance between regions as claimed. As noted above, Makram-
Ebeid's goal is to merge adjacent regions using the lowest value of the scale
parameter λ [Makram-Ebeid, col. 2, lines 8-10]. This teaching strongly suggests
that the lowest value of the scale parameter is the optimal scale parameter. And
such a scale parameter would also minimize variance between regions via
Makram-Ebeid's merging process that ultimately merges adjacent regions with
22
Page: Previous 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Next
Last modified: November 3, 2007