Appeal No. 2002-1491
Application 08/845,526
notwithstanding that according to the Sherman and Schulmeiss references the de Casteljau
process can be used to evaluate Bezier curves. The fact that the de Casteljau process can be used
to evaluate Bezier curves does not mean the NURBS representation of a curve and the Bezier
model of the same are equivalents.
As for the examiner’s statement that Jia also teaches generating a curve without first
converting the NURBS defined curve to a polygon mesh, we have the same problems with it as
earlier discussed in the context of the rejection of claims 1, 2, 6, 8, and 13-15.
In the Response to Argument portion of the examiner’s Answer, the examiner makes a
surprising statement. On page 14, in lines 15-17, the examiner states: “it is noted that Luken
teaches a graphics pipeline (Fig. 2) to render the parametric surface, and also teaches pipeline
to decompose NURBS to Bezier (Col. 1 45-50) [Emphasis added]”. That statement contradicts
the examiner’s earlier finding on page 10 of the Answer that “Luken fails to teach the use of
these methods [the various claimed steps] for a Bezier curve.” Upon closer scrutiny, we see that
there really is no contradiction. The examiner is only sloppy in referring to different parts of
Luken. Luken’s invention indeed is not described as using a Bezier curve or Bezier control
points, just as the examiner found on page 10 of the Answer. The examiner’s statement on page
14 of the Answer which appears to say the contrary actually refers not to Luken’s disclosed
invention but to a different invention referred to in the Background Art portion of Luken’s
specification, i.e., U.S. Patent No. 4,912,659. It is abundantly clear that the reference to Patent
No. 4,912,659 is describing something different from what is implemented in Luken’s own
10
Page: Previous 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Next
Last modified: November 3, 2007