Appeal No. 1999-2069
Application 08/397,639
three-dimensional body being imaged and that most pairs have
misregistration due to a combination of these sources (col. 1,
lines 64-68; Fig. 12). Kano utilizes nonlinear warping in order
to obtain improved registration between the two images (col. 5,
lines 60-64) which can be quite distorted (Figs. 9A&9B).
The Examiner finds that Smilansky and Frankot perform image
registration utilizing an affine transformation by calculating
the factors of affine transformation with the method of least
squares and then carrying out the affine transformation (FR5-6;
EA5-6). The Examiner concludes that it would have been obvious
to one of ordinary skill in the art to use an affine
transformation as taught by Smilansky for aligning the images in
Kano "since an affine transformation compensates for angular
deviation and scaling errors . . . and also considers translation
. . . and because it is very well known that the error on affine
transformations can be minimized using least squares" (FR5; EA5).
The Examiner concludes that it would have been obvious to use an
affine transformation as taught by Frankot for aligning the
images in Kano "since an affine transformation can be used to
optimize for minimizing MSE (Mean Squared Error) and reducing
computation" (FR6; EA6).
Smilansky discloses using an affine transformation to
register the image of a PCB (printed circuit board) being
inspected with a reference image (col. 6, line 41 to col. 7,
- 9 -
Page: Previous 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Next
Last modified: November 3, 2007