Ex Parte Kisilev et al

Patent Trial and Appeal Board

Feb 27, 2014

11481346 (P.T.A.B. Feb. 27, 2014)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte PAVEL KISILEV and SUK HWAN LIM
____________

Appeal 2011-010973
Application 11/481,346
Technology Center 2600
____________

Before BRUCE R. WINSOR, MICHAEL J. STRAUSS, and J. JOHN LEE,
Administrative Patent Judges.

WINSOR, Administrative Patent Judge.

DECISION ON APPEAL
Appellants1 appeal under 35 U.S.C. § 134(a) from a Final Rejection of
claims 1-20, which constitute all the claims pending in this application. We
have jurisdiction under 35 U.S.C. § 6(b).
We affirm-in-part and institute a new ground of rejection within the
provisions of 37 C.F.R. § 41.50(b) (2011).

1 The real party in interest identified by Appellants is Hewlett-Packard
Development Company, LP, a wholly-owned affiliate of Hewlett-Packard
Company, and having HPQ Holdings, LLC, as its general or managing
partner. App. Br. 2.
Appeal 2011-010973
Application 11/481,346

2

STATEMENT OF THE CASE
Appellants’ disclosure relates to “image processing techniques,
including . . . image analysis and image enhancement.” Spec. ¶ 01. Claim
1, which is illustrative, reads as follows:
1. A method of processing an image, comprising:
obtaining an image to be processed;
calculating via a central processing unit a classification
score for each of a plurality of locations in the image, each of
the classification scores indicating whether the corresponding
location primarily contains (i) background, (ii) at least one of
texture or weak edges, or (iii) strong features; and
differentially processing via the central processing unit
different areas of the image based on the classification scores to
generate a modified image, wherein each of the classification
scores is calculated by determining changes in pixel values
along a plurality of different directions.

The Examiner relies on the following prior art in rejecting the claims:
Elena Console & Marie Catherine Mouchot, New
Classification Techniques for Analysis of Remote Sensing
Integrated Data, 2 GEOSCIENCE & REMOTE SENSING 646
(1997).
Lily Rui Liang & Carl G. Looney, Competitive Fuzzy
Edge Detection, 3 APPLIED SOFT COMPUTING 123 (2003).
Tomoya Tokairin & Yoshiharu Sato, Design of
Nonlinear Adaptive Digital Filter by Wavelet Shrinkage, 86
ELECS. & COMM. JAPAN 59 (2003).
MathWorks.com, Foundations of Fuzzy Logic,
http://www.mathworks.com/access/helpdesk/help/toolbox/fuzz
y/bp7816_-1.html (archived July 19, 2006) [hereinafter
Mathworks].
Appeal 2011-010973
Application 11/481,346

3
Claims 1, 2, 6, 8-11, 13, 15, and 17-20 stand rejected under 35 U.S.C.
§ 102(b) as anticipated by Liang. Ans. 4-9.
Claims 3, 5, 12, and 14 stand rejected under 35 U.S.C. § 103(a) as
unpatentable over Liang and Tokairin. Ans. 9-11.
Claim 4 stands rejected under 35 U.S.C. § 103(a) as unpatentable over
Liang and Console. Ans. 11-12.
Claims 7 and 16 stand rejected under 35 U.S.C. § 103(a) as
unpatentable over Liang and Mathworks. Ans. 13.
Rather than repeat the arguments here, we refer to the Briefs (“App.
Br.” filed Dec. 6, 2010; “Reply Br.” filed Apr. 18, 2011) 2 and the Answer
(“Ans.” mailed Feb. 18, 2011) for the respective positions of Appellants and
the Examiner. Only those arguments actually made by Appellants have been
considered in this decision. Arguments that Appellants did not make in the
Briefs have not been considered and are deemed to be waived. See
37 C.F.R. § 41.37(c)(1)(vii).

ISSUES
The issues presented by Appellants’ contentions are as follows:
Does Liang disclose “differentially processing via the central
processing unit different areas of the image based on the classification scores
to generate a modified image,” as recited in claim 1?

2 The listing of claims in the Claims Appendix (App. Br. 15-18) contains
errors. In particular, the base claim for the majority of the dependent claims
is indicated to be “claim 0.” We refer to the text of the claims filed March
24, 2010, to ascertain the correct base claims for the dependent claims, but
refer to the text of the claims in the Claims Appendix as otherwise definitive
of the claims on appeal.
Appeal 2011-010973
Application 11/481,346

4
Does Liang disclose “differentially processing via the central
processing unit different areas of the image based on the measures of local
activity and the measures of local pixel-value variations to generate a
modified image,” as recited in claim 10?
Does Liang disclose “differentially processing via the central
processing unit the map of the fuzzy classification scores to generate a
modified image,” as recited in claim 18?
Does Liang disclose “displaying the map of the fuzzy classification
scores,” as recited in claim 19?
Does the combination of Liang and Tokairin teach or suggest “each of
the classification scores is used to set a threshold for determining whether to
modify coefficients corresponding to image data in the corresponding
location,” as recited in claim 3?
Does the combination of Liang and Tokairin teach or suggest “each
measure of local activity is used to set a threshold for determining whether
to modify coefficients corresponding to image data in the corresponding
region,” as recited in claim 14?
Does the combination of Liang and Mathworks teach or suggest “the
differential processing is a piece-wise function of data in the image, the
piece-wise function comprised of different segments determined by each of
the classification scores,” as recited in claim 7?
Does the combination of Liang and Mathworks teach or suggest “the
differential processing is a piece-wise function of data in the image, the
piece-wise function comprised of different segments determined by the
measures of local activity,” as recited in claim 16?

Appeal 2011-010973
Application 11/481,346

5
ANALYSIS
REJECTION UNDER 35 U.S.C. § 102(b)
Claim 1
Appellants contend as follows:
[W]hile this cited portion of Liang [(pp. 124-25 (Table 1; § 2.2,
ll. 1-17))], as well as the remaining portions of Liang, discloses
classification of regions of an image, there is no disclosure of
differentially processing the various regions of the image based
on the classification. That is, there is simply no disclosure in
Liang of processing regions of an image differently based upon
how the various regions have been classified.
App. Br. 7. Appellants further contend as follows:
[B]ecause the direction values in Liang appear to be generated
independent from the classification scores . . . any processing of
values based upon the direction values (such as the classes
disclosed by Liang) cannot be read as being analogous to
differentially processed values of “different areas of the image
based on the classification scores to generate a modified
image,” as recited in independent claim 1 . . . .
App. Br. 7-8.
We are not persuaded of error. As found by the Examiner (Ans. 4)
and conceded by Appellants (App. Br. 7), Liang discloses six prototype
feature vectors c0-c5 that are used to classify pixels of an image into classes,
Classes 0-5, based on the immediately adjacent pixels, i.e., “a classification
score for each of a plurality of locations in the image,” as recited in claim 1.
See Liang 124-25 (§§ 2.1-2; Figs. 1-3; Table 1). Liang’s direction values d1-
d4 (Liang 124) are not independent from the classification scores, but rather
make up Liang’s feature vector for the pixel that is compared to the
prototype vectors c0-c5 to classify the pixel into Classes 0-5. As explained
by the Examiner (Ans. 14-15), Liang further discloses edge pixel
Appeal 2011-010973
Application 11/481,346

6
modification and despeckling that processes the pixels (i.e., “different areas
of the image”) differently depending on the classification scores. See Liang
125-28 (§§ 2.4–3.3). For example, Liang discloses a pixel competition
process, which causes differential processing of different areas of the image
based on the classifications:
IF x is Class 0 (background) THEN change pixel to white.
IF x is Class 1 (edge) THEN
compete d3 with neighbor pixels in Direction 3
IF it wins THEN change it to black (edge) ELSE change
to white.
IF x is Class 2 (edge) THEN
compete d4 with neighbor pixels in Direction 4
IF it wins THEN change it to black (edge) ELSE change
to white.
IF x is Class 3 (edge) THEN
compete dl with neighbor pixels in Direction 1
IF it wins THEN change it to black (edge) ELSE change
to white.
IF x is Class 4 (edge) THEN
compete d2 with neighbor pixels in Direction 2
IF it wins THEN change it to black (edge) ELSE change
to white.
IF x is Class 5 (speckle edge) THEN change pixel to black
(edge).
Liang 127.
Appellants have failed to persuade us of error in the rejection of claim
1. Therefore, we sustain the rejection of claim 1.
Appeal 2011-010973
Application 11/481,346

7
Claim 10
Appellants present arguments regarding claim 10 (App. Br. 6-8) that
are substantially the same as those presented for claim 1 (id.), and are
unpersuasive for the reasons discussed above. Appellants further contend,
for the first time in the Reply Brief,
the Examiner has not shown any portion of Liang as disclosing
differentially processing different areas of an image based on
both measures of local activity and the measures of local pixel-
value variations, as recited in independent claim 10. That is,
the Examiner has made no effort to show any alleged disclosure
in Liang of differentially processing different areas of an image
based on a single fuzzy classification calculation may be read as
anticipating differentially processing an image based on
multiple calculations (e.g., the measures of local activity and
the measures of local pixel-value variations).
Reply Br. 2. Appellants’ argument is untimely and waived. See Ex parte
Borden, 93 USPQ2d 1473, 1474 (BPAI 2010) (informative) (“[T]he reply
brief [is not] an opportunity to make arguments that could have been made
in the principal brief on appeal to rebut the Examiner’s rejections, but were
not.”). Nevertheless, in the interest of administrative and judicial economy,
we note that Appellants’ additional argument is unpersuasive.
Appellants are correct (see Reply Br. 2) that the Examiner maps both
the “measures of local activity” (Ans. 6-7) and the “measures of local pixel-
value variation” (Ans. 7) to the classification of pixels resulting from
Liang’s fuzzy classification process. See Liang 124-25 (§§ 2.1-2; eqs. 1a-
2b; Figs. 1-3; Table 1). Appellants’ Specification explains that measures of
local activity are measures that are indicative of features in corresponding
local regions (Spec. ¶ 03) while measures of local pixel value variation are
measures that are indicative of an amount of variation in pixel values across
Appeal 2011-010973
Application 11/481,346

8
a corresponding local region (id.). We note that Claim 10 recites that
differentially processing different areas of the image is “based on the
measures of local activity and the measures of local pixel-value variations”
(emphasis added). It does not recite, however, how the differential
processing is based on the recited measures, nor does it recite that the
measures are applied in any particular order, are applied directly, or are
applied separately.
Liang’s classification process has a plurality of steps. Relevantly, a
feature vector is created by calculating distance values d1-d4 by comparing
the gray scale value of a pixel p5 with the gray scale values of neighboring
pixels p1-p4, p6-p9. Liang 124 (§ 2.1; Fig. 1: eqs. 1a-2b). The resulting
feature vector is indicative of an amount of variation in pixel values across a
local region (3x3 pixels), and is, therefore, a “measure[] of local pixel-value
variations,” as recited in claim 10. The feature vector is then compared to
the prototype vectors c0-c5 to classify the pixels into Classes 0-5, which
correspond to background and various edge classes. Liang 124-25 (§ 2.2;
Figs 2, 3; Table 1). The resulting Class is indicative of features (e.g., edges)
in the corresponding local area and is, therefore, a “measure[] of local
activity,” as recited in claim 10.
Liang discloses that the differential processing is based on the
classification of the pixel into Classes 0-5, i.e., the “measures of local
activity,” which is in turn based on the feature vector, i.e., the “measures of
local pixel-value variations.” Therefore, Liang discloses that the differential
processing is based on both “the measures of local activity and the measures
of local pixel-value variations,” as recited in claim 10.
Appeal 2011-010973
Application 11/481,346

9
Therefore, Appellants have not persuaded us of error in the rejection
of claim 10. Accordingly, we sustain the rejection of claim 10.
Claim 18
Appellants present arguments regarding claim 18 (App. Br. 6-8) that
are substantially the same as those presented for claim 1 (id.), and are
unpersuasive for the reasons discussed above. Appellants further contend,
for the first time in the Reply Brief, as follows:
[T]he Examiner has made no effort to show any alleged
disclosure in Liang of differentially processing different areas
of an image based on a map of the fuzzy classification scores.
Indeed, Appellants are unable to find any disclosure of Liang of
generating any map of fuzzy classification scores, let alone a
disclosure of differentially processing a map of the fuzzy
classification scores to generate a modified image.
Reply Br. 3. Appellants’ argument is untimely and waived. See Borden, 93
USPQ2d at 1474. Nevertheless, in the interest of administrative and judicial
economy, we note that Appellants’ additional argument is unpersuasive.
As discussed above, Liang discloses classifying individual pixels of
an image and using those classifications to differentially process areas of the
image to generate a modified image. It is inherent in Liang’s process that
the pixels, and thus their respective classifications, must necessarily be
identified by their individual locations in the original image, such as by
vertical and horizontal coordinates. Such necessary identification of the
pixels and their classification scores by location in the image inherently
discloses “a map of the fuzzy classification scores” (emphasis added) that is
used (i.e., “process[ed]”) to generate a modified image.
Therefore, Appellants have not persuaded us of error in the rejection
of claim 18. Accordingly, we sustain the rejection of claim 18.
Appeal 2011-010973
Application 11/481,346

10
Claim 19
The Examiner maps “displaying the map of the fuzzy classification
scores” to the illustrations at Table 1 and Figure 5 of Liang (Ans. 9, 17-18).
Appellants contend,
Table 1 cited by the Examiner appears to merely be a table in a
technical paper utilized to illustrate an example of the manner
in which classes and their prototype vectors may be grouped in
conjunction with the fuzzy classification method of Liang. See
Liang, Table 1. That is, there is no disclosure in Liang of any
actual display of the information contained in Table 1 of Liang
or even that this Table 1 may be read as a map of fuzzy
classification scores.
Reply Br. 6.
We agree with Appellants. The Examiner’s mapping of the
“displaying” step is to the presence of a table and a figure in the reference
itself, rather than to a disclosure that the technology described by the
reference performs the displaying step. However, the rejection of claim 18,
from which claim 19 depends, is based the technology described in Liang.
Therefore, the Examiner has failed to show that the elements relied on in
Liang are “arranged as in the claim under review,” In re Bond, 910 F.2d 831,
832 (Fed. Cir. 1990) (citation omitted). Furthermore, referring to claim 18,
we note that the map to be displayed in the displaying step of claim 18 is a
map of the scores calculated for each of a plurality of locations in the image
(e.g., each pixel). Liang’s Table 1 appears to be a map of a classification
scheme, not a map of calculated classification scores.
Appellants have persuaded us of error in the rejection of claim 19.
Accordingly, we do not sustain the rejection of claim 19.
Appeal 2011-010973
Application 11/481,346

11
Remaining Dependent Claims
Claims 2, 6, 8, 9, 11, 13, 15, 17, and 20, which depend, directly or
indirectly, from claim 1, 10, or 18, are not separately argued with
particularity. Accordingly, as with the independent claims from which they
depend, we sustain the rejection of claims 2, 6, 8, 9, 11, 13, 15, 17, and 20.

REJECTIONS UNDER 35 U.S.C. § 103(a)
Claim 3
The Examiner finds that Tokairin teaches or suggests “each of the
classification scores is used to set a threshold for determining whether to
modify coefficients corresponding to image data in the corresponding
location,” as recited in claim 3. Ans. 10, 18-19 (citing Tokairin, Abstract; p.
60-61 (§§ 3, 5)).
Appellants contend as follows:
Tokairin appears, at best, to merely teach removal of signal
nose [sic] in adaptive digital filters by shrinking empirical
wavelet coefficients in the wavelet domain (see Tokairin,
Summary). However, there is no teaching in Tokairin that
shrinking empirical wavelet coefficients is analogous to using
classification scores . . . to set a threshold, let alone to set a
threshold to determine “whether to modify coefficients
corresponding to image data,” as recited in claim[] 3.
Reply. Br. 4. We agree with Appellants.
Although we note that Tokairin does disclose the use of threshold
levels and adaptive filter coefficients, we find no mention in the passages
identified by the Examiner of use of Tokairin’s shrinking empirical wavelet
coefficients to process an image. The Examiner’s articulated rationale for
combining Tokairin with Liang, “allow[ing] user of Liang to classify image
and at the same time reducing noise introduce in the input signal by applying
Appeal 2011-010973
Application 11/481,346

12
the robust adaptive wavelet shrinking filter of Tokairin which update the
threshold levels adaptively” (Ans. 10), explains why it would be obvious to
apply Tokairin’s filter to an input signal to Liang’s classification. The
Examiner’s rationale does not explain, nor is it apparent from the passages
cited, why one of ordinary skill in the art would apply Tokairin’s filter to
“modify coefficients corresponding to image data in the corresponding
location” of an image, as recited in claim 3.
On this record, we find the Examiner has not established a prima facie
case of obviousness of claim 3. Therefore, Appellants have persuaded us of
error in the rejection of claim 3. Accordingly, we do not sustain the
rejection of claim 3.
Claim 14
For substantially the same reasons as set forth above regarding claim
3, we do not sustain the rejection of claim 14.
Claim 7
The Examiner finds that Mathworks, when combined with Liang,
teaches or suggests that “the differential processing is a piece-wise function
of data in the image, the piece-wise function comprised of different
segments determined by each of the classification scores,” as recited in
claim 7. Ans. 13 (citing Mathworks 5). More particularly, the Examiner
relies on Mathworks to demonstrate
that the fuzzy classifier of Liang inherently teaches a piece-
wise function, explained as follow [sic].
A fuzzy logic theory is based on membership functions,
and the membership functions are piece- wise functions. Thus,
the fuzzy classifier of Liang inherently teaches a piece- wise
function. The Examiner applied [Mathworks] just to show the
inherency, and the combination of Liang and [Mathworks] is
Appeal 2011-010973
Application 11/481,346

13
proper because the claims limitations describe the fundamental
properties of a fuzzy membership function which is well known
in the art.
Ans. 20 (emphases added).
Appellants contend “the cited portion of Mathworks, at best, appears
to be a general tutorial of fuzzy logic that includes a pair of illustrative
graphs that visually differentiate between two separate grouping possibilities
for a set of elements.” App. Br. 13. We agree with Appellants.
We agree with the Examiner that Mathworks teaches that fuzzy logic
is based on membership functions, and that membership functions may be
piece-wise, for example, Mathworks’ membership functions labeled trimf
and trapmf (Mathworks 5; see Ans. 20). However, we do not agree that
Mathworks shows that Liang inherently teaches piece-wise membership
functions.
“The inherent teaching of a prior art reference, a question of fact,
arises both in the context of anticipation and obviousness.” In re Napier, 55
F.3d 610, 613 (Fed. Cir. 1995). “[T]he examiner must provide a basis in fact
and/or technical reasoning to reasonably support the determination that the
allegedly inherent characteristic necessarily flows from the teachings of the
applied prior art.” Ex parte Levy, 17 USPQ2d 1461, 1464 (BPAI 1990)
(non-precedential). “Inherency . . . may not be established by probabilities
or possibilities. The mere fact that a certain thing may result from a given
set of circumstances is not sufficient.” In re Oelrich, 666 F.2d 578, 581
(CCPA 1981) (quoting Hansgirg v. Kemmer, 102 F.2d 212, 214 (CCPA
1939).
We find that Mathworks does not support the Examiner’s finding
(Ans. 13, 20) that Liang’s classification, i.e., membership, functions
Appeal 2011-010973
Application 11/481,346

14
necessarily must be piece-wise. See Levy, 17 USPQ2d at 1464. In addition
to membership functions that are piece-wise (Mathworks 5), Mathworks
discloses membership functions that are continuous, and not piece-wise, for
example, Mathworks’ membership functions labeled gaussmf, gauss2mf,
gbellmf, sigmf, dsigmf, psigmf, zmf, pimf, and smf (Mathworks 6).
On this record, we find the Examiner has not established a prima facie
case of obviousness of claim 7. Therefore, Appellants have persuaded us of
error in the rejection of claim 7. Accordingly, we do not sustain the
rejection of claim 7.
Claim 16
For substantially the same reasons as set forth above regarding claim
7, we do not sustain the rejection of claim 16.
Remaining Dependent Claims
Claims 4, 5, and 12, which depend, directly or indirectly, from claim 1
or 10, are not separately argued with particularity. Accordingly, as with the
independent claims from which they depend, we sustain the rejections of
claims 4, 5, and 12.

NEW GROUND OF REJECTION WITHIN 37 C.F.R. § 41.50(b)
Claims 3 and 14 are rejected on a new ground of rejection under
35 U.S.C. § 103(a) as unpatentable over Liang.
Claim 3
We adopt the Examiner’s findings and explanations regarding claim 1
(Ans. 4-5, 14-15) in light of our discussion of claim 1 above. We find that
Liang’s pixel competition process (Liang 127) teaches “the classification
scores [are] used to set a threshold for determining whether to modify
Appeal 2011-010973
Application 11/481,346

15
coefficients corresponding to image data in the corresponding location,” as
recited in claim 3.
Looking to the IF-THEN pairs of Liang’s pixel competition process
(id.), we find that “compet[ing]” the magnitude differences based on the
class of the pixel in order to determine how to change the pixel teaches
setting a threshold using the classification score, with the original gray value
of the pixel being modified to “black” or “white” being the “coefficient[]
corresponding to image data.” For example, consider the expression: “IF x
is Class 2 (edge) [(i.e., determined from the classification score)] THEN
compete d4 with neighbor pixels [(i.e., set a threshold)] in Direction 4 IF it
wins THEN change it to black (edge) ELSE change to white [(i.e., modify
coefficients corresponding to image data in the corresponding location)].”
Id.
Although Liang teaches the use of classification scores to set a
threshold, it does not teach that “each” of the classification scores is used to
set a threshold. In particular, Liang’s classification scores Class 0 and Class
5 are not used to set a threshold in the manner described above. However, it
would have been obvious to one of ordinary skill in the art to extend the
setting of thresholds to each classification score. To do so would be no more
than a duplication of elements, which has no patentable significance where,
as here, no new and unexpected result is produced. See In re Harza, 274
F.2d 669, 671 (CCPA 1960). Further, it is no more than a “combination of
familiar elements according to known methods . . . [that] does no more than
yield predictable results.” KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 416
(2007).
Appeal 2011-010973
Application 11/481,346

16
Claim 14
Claim 14 depends from independent claim 10. We adopt the
Examiner’s findings and explanations regarding claim 10 (Ans. 6-7, 15-16),
in light of our discussion of claims 1 and 10 above. We note that, as
discussed above regarding claim 10, Liang’s classification scores Class 0-5
are encompassed by the phrase “measures of local activity.” We find that
Liang teaches “each measure of local activity is used to set a threshold for
determining whether to modify coefficients corresponding to image data in
the corresponding region,” as recited in claim 14 for reasons that are
substantially the same as discussed above for claim 3.

DECISION
The decision of the Examiner to reject claims 1, 2, 4-6, 8-13, 15, 17,
18, and 20 is affirmed.
The decision of the Examiner to reject claims 3, 7, 14, 16, and 19 is
reversed.
We enter a new ground of rejection for claims 3 and 14 under
35 U.S.C. § 103(a).
This decision contains new grounds of rejection pursuant to 37 C.F.R.
§ 41.50(b). Section 41.50(b) provides that “[a] new ground of rejection . . .
shall not be considered final for judicial review.”
Section 41.50(b) also provides that Appellants, WITHIN TWO
MONTHS FROM THE DATE OF THE DECISION, must exercise one of
the following two options with respect to the new ground of rejection to
avoid termination of the appeal as to the rejected claims:
Appeal 2011-010973
Application 11/481,346

17
(1) Reopen prosecution. Submit an appropriate
amendment of the claims so rejected or new evidence relating
to the claims so rejected, or both, and have the matter
reconsidered by the examiner, in which event the proceeding
will be remanded to the examiner. . . .
(2) Request rehearing. Request that the proceeding be
reheard under § 41.52 by the Board upon the same record.
37 C.F.R. § 41.50(b).
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1). See 37 C.F.R.
§ 1.136(a)(1)(iv).

AFFIRMED-IN-PART
37 C.F.R. § 41.50(b)

bab