Ex Parte Luo - Page 22


                   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