Ex Parte Fan

Board of Patent Appeals and Interferences

Dec 16, 2011

11101869 (B.P.A.I. Dec. 16, 2011)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte ZHIGANG FAN
____________

Appeal 2009-014748
Application 11/101,869
Technology Center 2600
____________

Before SALLY C. MEDLEY, ALLEN R. MacDONALD,
and DENISE M. POTHIER, Administrative Patent Judges.

POTHIER, Administrative Patent Judge.

DECISION ON APPEAL

Appeal 2009-014748
Application 11/101,869

2
STATEMENT OF THE CASE
Appellant appeals under 35 U.S.C. § 134(a) from the Examiner’s
rejection of claims 1-20. We have jurisdiction under 35 U.S.C. § 6(b). We
reverse.

Invention
Appellant’s invention relates to error diffusion methods for halftoning
images. See Spec. ¶ 0001; Abstract. Claim 1 is reproduced below with the
key disputed limitations emphasized:
1. Apparatus comprising:
a color separation processor to separate a source image having a
given color scheme into plural color separations of a different color
scheme, the plural color separations including at least one separation
corresponding to one color component and another separation
corresponding to another color component different than the one color
component;
an error diffusion engine;
a controller to control the error diffusion engine to carry out error
diffusion processing on the one separation based on error diffusion
processing on the another separation, the error diffusion processing on
the one separation including determining, for the one separation, a
given rendered pixel value of the one separation and corresponding to
a given pixel location, the given rendered pixel value being
determined by applying a quantization function to a pre-rendered
pixel value of the one separation and corresponding to the given pixel
location, the quantization function, when applied, performing
mathematical operations on each of (i) a modified input pixel value of
the one separation and corresponding to the given pixel location, the
modified input pixel value being determined as a function of an error
value of a pre-rendered pixel value of the one separation, and (ii)
another error value, the another error value being determined as a
function of rendered pixel values of the another separation; and
Appeal 2009-014748
Application 11/101,869

3
rendering circuitry to output a rendered image including the same
plural color separations, the output rendered image being populated
with, for the one separation, the rendered pixel value determined by
applying the quantization function.

The Examiner relies on the following as evidence of unpatentability:
Harrington US 5,631,748 May 20, 1997
Quintana US 2004/0100646 A1 May 27, 2004
(filed Nov. 25, 2002)

THE REJECTION
The Examiner rejected claims 1-20 under 35 U.S.C. § 103(a) as
unpatentable over Harrington and Quintana. Ans. 3-15.1

THE CONTENTIONS
Independent claim 1 recites a controller to control an error diffusion
engine that carries out error diffusion processing on one separation based on
the error diffusion processing on another separation. Claim 1 defines the
error diffusion processing on the one separation to include determining a
rendered pixel value of the one separation and corresponding to a given
pixel location, the value being determined by applying a quantization
function to a pre-defined pixel value of the one separation and corresponding
to the given pixel location. The quantization function in claim 1 is further
defined as, when applied, performing mathematical operations on each of (i)
a modified input pixel value of the one separation and corresponding to the

1 Throughout this opinion, we refer to (1) the Appeal Brief filed March 2,
2009 and supplemented November 21, 2011 and (2) the Examiner’s Answer
mailed May 20, 2009.
Appeal 2009-014748
Application 11/101,869

4
pixel location, the modified input pixel value being determined as a function
of an error value of a pre-defined pixel value of the one separation, and
(ii) another error value, the another error being determined as a function of
rendered pixel values of the another separation. Appellant discusses what
portions of disclosure support these limitations. See Supp. App. Br. 13-19.
With specific regard to these limitations, the Examiner finds that
Harrington teaches in various columns a controller to process one separation
based on processing another separation by determining a given rendered
pixel value through applying a quantization function to a pre-rendered pixel
value. Ans. 3-4. Specifically, citing to Figure 5 and various columns (e.g.,
columns 5-7) in Harrington, the Examiner states Harrington’s halftoning
process that generates halftone screens includes adding a magnitude (e.g., an
error) “of [the] previously placed colors to the current color being placed on
each color separation based on a previous color component and
corresponding to the pixel position[.]” Ans. 4 (citations omitted). The
Examiner relies on Quintana to teach the recited error diffusion engine and
an error diffusion technique. See Ans. 5, 15.
Among other arguments, Appellant contends that Harrington fails to
teach the error diffusion processing limitations and, more specifically, the
quantization function recited in claim 1. App. Br. 16; Supp. App. Br. 23.
The Board requested additional briefing to state clearly: (1) what the
recited quantization function encompasses, the mathematical operations that
function performs, and the results of the operations; (2) what the recited
pre-rendered pixel value is and how the value is applied to the quantization
function; (3) how the recited modified input pixel value is determined as a
function of an error value; and (4) how the claimed “error diffusion
Appeal 2009-014748
Application 11/101,869

5
processing” differs from the cited prior art and the known concept of error
diffusion. See ORDER – 37 C.F.R. § 41.50(d) mailed October 21, 2011. In
response, Appellant submitted a Supplemental Appeal Brief along with a
Declaration from the Inventor Fan in support of the Supplemental Appeal
Brief. See Supp. App. Br. 13-24 and Appendix A.

ISSUE
Under § 103, has the Examiner erred by finding that Harrington and
Quintana collectively would have taught or suggested a controller to control
the error diffusion engine to carry out error diffusion processing, including
determining a given rendered pixel value of the one separation and
corresponding to give pixel location, the value being determined by applying
a quantization function to a pre-defined pixel value of the one separation that
performs mathematical operations on each of (i) a modified input pixel value
of the one separation, the modified input pixel value being determined as a
function of an error value of a pre-defined pixel value of the one separation
and (ii) another error value being determined as a function of rendered
values of the another separation?

ANALYSIS
We find error in the Examiner’s rejection of claim 1, which calls for a
controller to control an engine diffusion engine that processes one separation
based on error diffusion processing on another separation, including
determining a rendered pixel value of the one separation by applying a
quantization function as recited. At the outset, we note that we have not
considered the Declaration submitted with the Supplemental Appeal Brief.
Appeal 2009-014748
Application 11/101,869

6
Also, we agree with the Examiner that Harrington and Quintana collectively
teach using an error diffusion technique on one separation and find a
motivation for combining the references. See Ans. 5, 15-16. That is, both
are in the same field of endeavor or halftoning processes (Ans. 15-16), and
both are concerned with generating separations while avoiding overlapping
(see Ans. 16).
Also, contrary to Appellant’s assertions (App. Br. 17-23; Supp. App.
Br. 24-26), the Examiner does not propose a wholesale replacement of
Harrington’s process with Quintana’s error diffusion technique. Instead, the
Examiner proposes to make Harrington’s technique, including its adding and
comparing functions, into an error diffusion process with error values or
adjustments (e.g., adding a magnitude). See Ans. 5, 15-17. As the Examiner
indicates (Ans. 16), “[t]he test for obviousness is not whether the features of
a secondary reference may be bodily incorporated into the structure of the
primary reference . . . Rather, the test is what the combined teachings of the
references would have suggested to those of ordinary skill in the art.” In re
Keller, 642 F.2d 413, 425 (CCPA 1981). Moreover, Harrington does not
state that it uses a dithering technique. App. Br. 15; Supp. App. Br. 22.
Rather, Appellant points to a discussion in the Background of the Invention
section addressing Holladay (U.S. Patent No. 4,149,194). See Harrington,
col. 1, ll. 36-48. We therefore are not persuaded by Appellant’s arguments
asserting that Harrington’s system is not combinable or is incompatible with
Quintana’s error diffusion technique or vice versa. App. Br. 15-23; Supp.
App. Br. 22-26.
However, based on the record before us, we do not find that
Harrington teaches or suggests all the specific recitations of the quantization
Appeal 2009-014748
Application 11/101,869

7
function as recited, including performing mathematical operations on each
of (1) a modified input pixel value of the one separation, the modified input
pixel value being determined as a function of an error value of a pre-defined
pixel value of the one separation and (2) another error value being
determined as a function of rendered values of the another separation. The
Examiner relies on Figure 5 and its accompanying descriptions to teach this
function and maps the discussed magnitude to the recited error values. See
Ans. 4.
Harrington teaches a halftoning process where separation image
signals (e.g., Kp, Mp, Cp, Yp) are received by the halftoning system at inputs
110, 112, 114, and 116. Col. 6, ll. 40-44; Fig. 5. Harrington’s image signals
are colorant values and represent the relative amounts of black, magenta,
cyan, and yellow toners that are to be deposited over a given area. Col. 4, ll.
36-42. Harrington also teaches that the separation image signals are added
together at signal adders (e.g., 135, 137) to yield other values (e.g., Mp’, Cp’).
Col. 6, l. 66 – col. 7, l. 3; col. 7, ll. 12-14; Fig. 5. For example, image
signals Kp and Mp are added together at adder 135 and serves to potentially
increase the magnitude of Mp’ or create a modified input signal for the
magenta toner. Col. 6, l. 66 – col. 7, l. 3; Fig. 5.
Harrington further discusses that a threshold value (e.g., t(i,j) or t in
Fig. 5) for a given location i,j and pixel in the image are used and compared
with separation image signals. Col. 6, ll. 49-60; Fig. 5. For example, the
modified input signal, Mp’, is compared with a threshold (e.g., t) for a given
location and pixel in the image to determine Mq and ultimately Msep.
Col. 6, ll. 10-31; col. 6, l. 66- col. 7, l. 11; Fig. 5. Harrington thus teaches
applying a function that performs mathematical operations on a modified
Appeal 2009-014748
Application 11/101,869

8
input pixel value (e.g., Mp’) of one separation (e.g., magenta separation) and
corresponding to the given pixel location. See id. Additionally, this value,
Mp’, is arguably determined as a function of an error value (e.g., Kp) or a
magnitude of previously placed colors to the color currently being placed
(col. 5, ll. 50-56) as the Examiner explains. See Ans. 4, 17. These
previously placed colors to the color current being placed collectively
represent another separation (e.g. the black separation) as recited.
However, the Examiner does not illustrate nor can we find in
Harrington a teaching or suggestion of performing operations on the
modified input pixel value (e.g., Mp’), which is determined as a function of
an error value of a pre-rendered pixel value of the one separation (e.g., the
magenta separation) as recited. Rather, the proposed error value (e.g., Kp) is
a value of previously placed colors or values of another separation (e.g.,
black separation). This is also true for the other separations, such as the
cyan separation (Cp), where the added or subtracting magnitudes (e.g., Kp +
Mp) are for other separations and not the one separation (e.g., Cp). See
Harrington, Fig. 5. We therefore agree with Appellant (App. Br. 16) that
Harrington fails to teach or suggest the specific quantization function recited
in claim 1, including the performing mathematical operations on a modified
input pixel value, which is determined as a function of an error value of a
pre-rendered pixel value of the one separation.
Based on the evidence of record, Appellant has persuaded us of error
in the obviousness rejection of: (1) independent claim 1; (2) independent
claims 7 and 17 which recite commensurate limitations; and (3) dependent
claims 2-6, 8-16, and 18-20 for similar reasons.

Appeal 2009-014748
Application 11/101,869

9
CONCLUSION
The Examiner erred in rejecting claims 1-20 under § 103.

DECISION
The Examiner’s decision rejecting claims 1-20 is reversed.

REVERSED

msc