Appeal No. 1997-4067
Application 08/285,324
pixel is present in a photograph region, error diffusion is
used, and when a pixel is present in a text region, the
image signal is quantized by a predetermined threshold value
(col. 8, lines 1-11) (because the compensation error (CE) is
zero).
The Examiner's position is that the weighting
coefficients from the weighting coefficient memory 7
multiplied by the compensation error signal (CE), which
changes depending on the image region type according to the
average maximum density differential signal (AD), constitute
first and second sets of weighting coefficients (FR10-11;
EA8). The Examiner concludes that it would have been
obvious to apply Kanno's method in the segmentation
processing of Williams.
Since Kanno shows segmentation of an image into text
and photograph portions, it is not clear why the Examiner
feels it is necessary to combine Kanno with Williams.
Nevertheless, the combination of references is not the
problem.
Appellants argue (Br18-19): "[Kanno's] process of
correcting the error does not affect how the error is
- 12 -
Page: Previous 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Next
Last modified: November 3, 2007