Appeal No. 2003-1816 Page 2
Application No. 09/195,593
Instead of storing (M×N) × (H×K) threshold values for the screen, the appellant stores
three sets of data. The first set of data, T(IP, JP), represents H×K thresholds of the base
halftone cell where (IP, JP) are the coordinates of each pixel of the base cell. The
second set of data, C(IP, JP), represents H×K increments of thresholds corresponding to
the first set of data. For example, if a threshold value in the first set represents the level
corresponding to k-pixel pattern, Tk, the increment in the second set is Ck = Tk -Tk-1. The
third set of data, F(IC, Jc), represents M×N fill-in indices, which address the cells, where
(IC, Jc) are the coordinates of "referring" pixel of the specific cell in the composite
screen. (Id. at 33-34.)
The thresholds of the first set and the increments of the second set are
represented in an eight-bit format. The indices of the third set are represented in a
format of no more than eight bits. Accordingly, no more than 2*(H×K) + (M×N) bytes of
memory are needed to store the screen. To store a 192×192 stochastic-cluster screen,
for example, the appellant explains that "four bits can be enough to specify the fill-in
order. . . ." (Id. at 34.) They assert that storing the screen in a conventional halftone
format would require 36,864 bytes of memory. In contrast, storing the screen in the
appellant's format would require only 2*18 + 1024 = 1060 bytes of memory. (Id.)
Page: Previous 1 2 3 4 5 6 7 8 9 Next
Last modified: November 3, 2007