Appeal No. 1997-4067
Application 08/285,324
line 3). Thus, Williams discloses step (a) and discloses
steps (b) and (c) except for the last limitation of
"generating an error and binary value therefrom." Williams
also fails to disclose steps (d) and (e).
The Examiner relies on Kanno. Kanno discloses
binarizing a multi-level pixel (B output at 41), generating
an error therefrom (EB), and using error diffusion. Kanno
notes that a maximum density difference of pixels within a
predetermined 4x4 area surrounding the pixel of interest is
large for a character (text) region having a high contrast
and is small for a photograph region having a gray scale
(col. 2, lines 38-42). Kanno uses the average maximum
density difference signal (AD) to adjust quantization of the
pixel of interest to either the quantization error (EB) (in
a photograph region) or zero (in a text region) (col. 5,
lines 55-60). When the compensation error signal (CE) is
not zero (i.e., for a photograph region), weighting
coefficients from the weighting coefficient memory 7 are
used to distribute the error to surrounding pixels (col. 6,
lines 12-30). The weighting coefficient memory 7 contains
only a single set of weighting coefficients. Thus, when a
- 11 -
Page: Previous 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Next
Last modified: November 3, 2007