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 22Page: Previous 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 NextLast modified: November 3, 2007