Appeal No. 2004-2297 Page 5
Application No. 10/196,817
1468, 1472, 223 USPQ 785, 788 (Fed. Cir. 1984); and In re
Rinehart, 531 F.2d 1048, 1052, 189 USPQ 143, 147 (CCPA 1976).
The examiner's position (answer, page 4) is that Brailean
does not explicitly disclose distinguishing both foreground and
background in the disclosed comparison of edge locations in
successive field periods. To overcome this deficiency of
Brailean, the examiner turns to Rosenberg for a teaching of
detecting objects in an image where edge locations in successive
frame periods are compared to distinguish both foreground and
background. The examiner argues (id.) that “[b]y comparing edges
in successive frames, pixels that are not varied by intensity by
more than a threshold within a given number of frames are
included as background, otherwise as foreground." In the
examiner's opinion (id.) the modification would have been obvious
because “because obtaining exact contours in an image will result
in a reliable motion estimation and therefore will achieve very
good estimates of both the displacement, as well as, background
and foreground of the moving objects within a video sequence.”
Appellants acknowledge (brief, page 4) that while Rosenberg
does detect edges, Rosenberg does not detect edges in the motion
vector field. It is argued (id.) that Rosenberg does not compare
edge locations in successive field periods, but rather only
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 Next
Last modified: November 3, 2007