6959293 et al.

Patent Trials and Appeals Board

Aug 29, 2019

90014056 - (D) (P.T.A.B. Aug. 29, 2019)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
90/014,056 12/15/2017 6959293 0716030.121-US2 1361
155332 7590 08/29/2019
Byrne Poh LLP / IPT
11 Broadway
Suite 760
New York, NY 10004
EXAMINER
BANANKHAH, MAJID A
ART UNIT PAPER NUMBER
3992
MAIL DATE DELIVERY MODE
08/29/2019 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte IMAGE PROCESSING TECHNOLOGIES, LLC
Patent Owner and Appellant
____________

Appeal 2019-004305
Reexamination Control 90/014,056
United States Patent 6,959,293 B2
Technology Center 3900
____________

Before JOHN A. JEFFERY, JONI Y. CHANG, and
JENNIFER L. MCKEOWN, Administrative Patent Judges.

JEFFERY, Administrative Patent Judge.

Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

2
DECISION ON APPEAL
Appellant appeals under 35 U.S.C. §§ 134 and 306 the Examiner’s
decision to reject claim 1.1 We have jurisdiction under 35 U.S.C. §§ 134
and 306, and we heard the appeal on August 1, 2019.
We REVERSE.

STATEMENT OF THE CASE
This proceeding arose from a request for ex parte reexamination filed
on December 15, 2017, of United States Patent 6,959,293 (“’293 patent”),
issued to Patrick Pirim, on October 25, 2005.
The ’293 patent describes a visual perception processor with
histogram calculation units that (1) receive data via a single data bus, and (2)
supply classification information to a time coincidences bus. In a preferred
embodiment, the histogram calculation units are organized into a matrix.
See generally ’293 patent, Abstract. Claim 1 is illustrative of the invention
and reproduced below:
1. A visual perception processor for automatically
detecting an event occurring in a multidimensional space
(i, j) evolving over time with respect to at least one
digitized parameter in the form of a digital signal on a data
bus, said digital signal being in the form of a succession
aijT of binary numbers associated with synchronization
signals enabling to define a given instant (T) of the
multidimensional space and the position (i, j) in this space,
the visual perception processor comprising:

the data bus;

1 Appellant identifies Image Processing Technologies, LLC as the real party
in interest. App. Br. 5.
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

3

a control unit;

a time coincidences bus carrying at least a time
coincidence signal; and

at least two histogram calculation units for the treatment of
the at least one parameter, the histogram calculation units
being configured to form a histogram representative of the
parameter as a function of a validation signal and to
determine by classification a binary classification signal
resulting from a comparison of the parameter and a
selection criterion C, wherein the classification signal is
sent to the time coincidences bus, and wherein the
validation signal is produced from time coincidences
signals from the time coincidence bus so that the
calculation of the histogram depends on the classification
signals carried by the time coincidence bus.

RELATED PROCEEDINGS
This appeal is said to be related to various proceedings, namely: (1)
two district court cases, one of which is said to be currently pending; and (2)
two inter partes review proceedings. App. Br. 5.2 In one cited inter partes
review proceeding, Samsung Electronics Co., Ltd. v. Image Processing
Technologies LLC, IPR2017-00336 (PTAB May 9, 2018) (“’336 IPR”),
another panel of this Board held, among other things, that claim 1 of the
’293 patent was not shown to be unpatentable as obvious over Pirim—a

2 Throughout this opinion, we refer to: (1) the Final Office Action, mailed
September 7, 2018 (“Final Act.”); (2) the Appeal Brief, filed January 7,
2019 (“App. Br.”); (3) the Examiner’s Answer mailed February 12, 2019
(“Ans.”); and (4) the Reply Brief filed April 11, 2019 (“Reply Br.”).

Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

4
prior art reference at issue here—combined with various other prior art
references that are not at issue here. See ’336 IPR, 10–76.
In the other cited inter partes review proceeding, Samsung Electronics
Co., Ltd. v. Image Processing Technologies LLC, IPR2017-01189 (PTAB
Aug. 18, 2017) (“’1189 IPR”), another panel of this Board denied institution
of inter partes review because the Petitioner did not demonstrate a
reasonable likelihood that various claims of the ’293 patent other than claim
1 were unpatentable as obvious over either Pirim alone, or Pirim combined
with various other prior art references that are not at issue here. See ’1189
IPR, 8–27.

THE REJECTIONS
The Examiner rejected claim 1 under 35 U.S.C. § 103 as unpatentable
over Pirim (WO 99/36893 A1; published July 22, 1999) and Howard Jay
Siegel et al., PASM: A Partitionable SIMD/MIMD System for Image
Processing and Pattern Recognition, 30 IEEE Trans. on Computers 934–45
(1981) (“Siegel”). Final Act. 9–15.
The Examiner rejected claim 1 under 35 U.S.C. § 103 as unpatentable
over Pirim and Hirota (US 6,118,895; issued Sept. 12, 2000). Final Act. 15–
18.

THE OBVIOUSNESS REJECTION OVER PIRIM AND SIEGEL
Regarding independent claim 1, the Examiner finds that Pirim
discloses, among other things, a visual perception processor with at least two
histogram calculation units, namely histogram formation blocks 24–29 in
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

5
Figure 12, for treating at least one parameter, namely “SR,” “V,” “VL,”
“DI,” and “x(m)1,” and “y(m)1.” Final Act. 11–15; Ans. 48–51. According
to the Examiner, these units are configured to: (1) form a histogram
representative of the parameter as a function of a validation signal; and (2)
determine by classification a binary classification signal resulting from
comparing the parameter and a selection criteria, where (a) the classification
signal is sent to a time coincidences bus 23, and (b) the validation signal is
produced from time coincidences signals from that bus so that the
histogram’s calculation depends on the classification signals carried by the
bus. Final Act. 12–15; Ans. 51
Despite finding that Pirim discloses every element of claim 1, the
Examiner nonetheless cites Siegel “to the extent that the Patent Owner . . .
argues that the portion at 37 in Pirim PCT fails to disclose two histogram
calculation units treating the same parameter because the axes are rotated.”
Final Act. 12–13. According to the Examiner, Siegel teaches parallel
processors that each process and form a histogram of the same parameter,
and, in light of this teaching, concludes that the claim would have been
obvious. Id. at 12–15; Ans. 51.
Appellant argues, among other things, that not only is the Examiner’s
claim construction under 35 U.S.C. § 112, paragraph six, erroneous, the
claim is patentable over the cited prior art in any event. App. Br. 10–17; see
also Reply Br. 12–13. Appellant also contends that the Examiner not only
improperly interprets the recited two histogram calculation units, the
Examiner’s reliance on Pirim is also said to be misplaced because Pirim
does not disclose two histogram calculation units configured to form a
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

6
histogram of a common parameter, but rather these units are said to process
a different parameter—a fact emphasized by the Board in the ’336 IPR
proceeding. App. Br. 33–44; Reply Br. 14–16, 24–27. Appellant adds that
not only is there no motivation to combine Pirim and Siegel as the Examiner
proposes, the proposed combination is inoperable in any event. App. Br.
48–57; Reply Br. 18–24, 27–29.

ISSUES
(1) Under 35 U.S.C. § 103, has the Examiner erred in rejecting claim
1 by finding that Pirim and Siegel collectively would have taught or
suggested a visual perception processor with at least two histogram
calculation units for treating at least one digitized parameter in the form of a
digital signal on a data bus, where the units are configured to (1) form a
histogram representative of the parameter as a function of a validation
signal, and (2) determine by classification a binary classification signal
resulting from comparing the parameter and a selection criteria, where (a)
the classification signal is sent to a time coincidences bus, and (b) the
validation signal is produced from time coincidences signals from that bus
so that the histogram’s calculation depends on the classification signals
carried by the time coincidences bus?
(2) Is the Examiner’s combining the teachings of these references
supported by articulated reasoning with some rational underpinning to
justify the Examiner’s obviousness conclusion?

Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

7
ANALYSIS
As noted previously, the patentability of claim 1 of the ’293 patent
was at issue in the related ’336 IPR proceeding. There, another panel of this
Board held, among other things, that claim 1 of the ’293 patent was not
shown to be unpatentable as obvious over Pirim—the very reference at issue
here—but combined with various other prior art references that are not at
issue here. See ’336 IPR 10–76.
That is, the earlier panel did not address the particular obviousness
issue here that is based on Pirim combined with either Siegel or Hirota, for
the latter two references were not before the earlier panel in the related IPR
proceedings. Nevertheless, the panel’s construction of claim 1 and its
findings and conclusions regarding Pirim in the ’336 IPR are relevant here.
In the ’336 IPR, the earlier panel construed “the [at least two]
histogram calculation units being configured to form a histogram
representative of the parameter” as recited in claim 1 as “the at least two
histogram calculation units being configured to each form a histogram
representative of at least one common parameter.” See ’336 IPR 14
(emphasis in original). In arriving at this construction, the earlier panel
noted the 16 polyvalent histogram calculation units 1a00 to 1a33 in the ’293
patent’s Figure 32, where each unit has access to various parameters,
including luminance “L,” tone “T,” saturation “S,” speed “V,” and direction
“D,” via bus 510. Id. at 13 (citing ’293 patent, col. 21, ll. 37–63). As the
earlier panel indicated, each histogram calculation unit can not only be
timeshared among different parameters during each frame, but also calculate
histograms and associated statistics for two or more parameters. See ’336
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

8
IPR 13 (citing ’293 patent, col. 21, ll. 18–36). The panel added that control
unit 513 determines which parameter to process at a given time by one or
several of the histogram calculation units. See ’336 IPR 13–14 (citing ’293
patent, col. 21, ll. 42–47).
Notably, the panel in the ’336 IPR found that Pirim does not disclose
at least two histogram calculation units being configured to each form a
histogram representative of at least one common parameter consistent with
the panel’s construction. See ’336 IPR 46. In reaching this finding, the
panel emphasized that Pirim’s Figure 12 shows each histogram formation
block 24 to 29 processing a different parameter—not the same parameter.
Id. That is, the panel found that each histogram formation block 24 to 29
processes a different parameter, namely luminance, speed, direction, time
constant, and x and y positions, respectively, such that no two blocks are
configured to process the same parameter as required by claim 1. Id. at 47.3
Given these findings, the Examiner’s reliance on Pirim is problematic
on this record. Although Pirim’s Figure 12 shows histogram calculation
units 24 to 29 that each treat their respective parameters, they are not
configured to each form a histogram representative of at least one common
parameter consistent with the earlier panel’s construction noted above. See
’336 IPR, at 46–47. Accord App. Br. 46–48; Reply Br. 25–26 (noting this

3 Notably, the earlier panel’s findings in this regard were made despite
histogram formation blocks 25 and 26 both receiving binary validation
signal “VL” in Pirim’s Figure 12—the only such blocks receiving that
signal. See Pirim 23 (noting that binary validation signal “VL” indicates
whether the result of the speed and oriented direction is valid).
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

9
point); Ans. 45 (acknowledging that each processing block in Pirim’s Figure
12 processes a different parameter).
To the extent the Examiner finds that the functionality associated with
Pirim’s Figures 15A and 15B teaches multiple histogram calculation units
configured to treat a common parameter, ostensibly by simultaneously
processing a parameter such as a pixel’s x-position as the Examiner seems to
suggest (see Final Act. 12; Ans. 50), such a finding is problematic on this
record, for it conflicts with the earlier panel’s findings in this regard. As the
earlier panel indicated, each histogram formation unit in Pirim is configured
to form a histogram for a different rotated axis—not the same axis. See ’336
IPR 47–48 (citing Pirim 37; Figs. 15A–B). Given these facts, the earlier
panel held that the Petitioner in that case did not show that Pirim would have
taught or suggested plural histogram calculation units configured to form a
histogram representative of at least one common parameter as claim 1
requires. See ’336 IPR 48.
Apparently recognizing this deficiency, the Examiner nonetheless
cites Siegel “to the extent that the Patent Owner . . . argues that the portion at
37 in Pirim PCT fails to disclose two histogram calculation units treating the
same parameter because the axes are rotated.” Final Act. 12–13. According
to the Examiner, Siegel teaches parallel processors that each process and
form a histogram of the same parameter, and, in light of this teaching,
concludes that the claim would have been obvious. Id. 12–15; Ans. 51. In
reaching this conclusion, the Examiner reasons that, in view of Siegel, it
would have been obvious to add another undisclosed element, which the
Examiner labels “28a,” that would be adjacent and similar to Pirim’s
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

10
histogram formation block 28 in Figure 12, so that both blocks would
process the same parameter, namely x-position, for different pixel segments.
Ans. 50–51. Alternatively, the Examiner reasons that, in light of Siegel, it
would have also been obvious for ordinarily skilled artisans to simply
configure two of Pirim’s existing histogram calculation units to treat the
same parameter to increase processing speed. Ans. 51.
This position is problematic on this record. To be sure, parallel
processing is a well-known technique that uses multiple processors to
execute multiple computer operations simultaneously. See Steven M.
Kaplan, WILEY ELECTRICAL & ELECTRONICS ENGINEERING DICTIONARY 552
(2004) (defining “parallel processing”); see also SAP Am. v. InvestPic, LLC,
898 F.3d 1161, 1170 (Fed. Cir. 2018) (holding that recited parallel
processing computer architectures did not add significantly more to the
abstract idea in that case because the parallel processing architectures were
no different than those in existing systems).
But as Dr. Bovik explains, adding histogram units to Pirim’s circuit in
Figure 12 as the Examiner proposes would not increase processing speed
because the histograms would be completed almost instantaneously as the
last piece of data is received in the histogram units. See Alan Conrad Bovik
Declaration Under 37 C.F.R. § 1.132, dated June 26, 2018, ¶ 20 (“Bovik
Decl.”). That is, even with multiple histogram units, each unit would still
operate on the same data at the same speed as Dr. Bovik indicates. Id.
We likewise find problematic the Examiner’s proposal to add another
undisclosed element, which the Examiner labels “28a,” that would be
adjacent and similar to Pirim’s histogram formation block 28 in Figure 12,
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

11
so that both blocks would process the same parameter, namely x-position,
for different pixel segments. See Ans. 50–51. Notably, this modification
would not accelerate histogram calculation as the Examiner seems to suggest
(see Final Act. 13), but would rather slow it down by, among other things,
combining the units’ respective histograms together after receiving the last
pixel. Bovik Decl. ¶ 21. As Dr. Bovik explains, this process would require,
among other things: (1) reading each histogram’s values out of memory; (2)
adding these values together; and (3) writing them back into memory. Id.
These slowdowns resulting from this requisite histogram aggregation
undercut the Examiner’s speed-enhancement rationale for the proposed
combination.
Also, the Examiner’s proposed modification would generate
validation signals based on classifications of different pixels—an anomalous
result given the purpose of Pirim’s validation signal, namely to determine
whether a single pixel should be included in a histogram as Dr. Bovik
indicates. Id. ¶¶ 23–24. To suggest that Dr. Bovik’s declaration is somehow
“incorrect” and based on a “wrong assumption” as the Examiner asserts
(Ans. 53–54) ignores the fact that the Examiner’s proposed combination is
premised on using data from different pixel segments to process the same
parameter. See Final Act. 13; Ans. 51. As Appellant indicates, under the
proposed combination, this data would be unrelated, at least with respect to
its respective regions. Reply Br. 27–29. Therefore, the proposed
combination—not Pirim itself—would use a classification signal from a
different pixel to decide whether to include a current pixel in a histogram,
which is an anomalous result at best. See id.; see also Bovik Decl. ¶¶ 23–24.
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

12
To the extent the Examiner finds otherwise, the evidence on this record does
not substantiate such a finding.
As Dr. Bovik explains, using the same two classification signals for
each of two histogram calculation units in the present invention is important
because it allows the classification results from the two units to be accounted
for when the data associated with each pixel is evaluated for addition to the
two histograms. Bovik Decl. ¶ 14. Notably, the claimed invention enables
evaluating data from two different perspectives simultaneously. Id. For
example, the claimed invention allows for simultaneously calculating (1)
one histogram for pixels having, for example, a strong red and weak green
component, and (2) another histogram for pixels having a strong red
component and a strong green component. Id.
To be sure, the Examiner’s maps, among other things, Pirim’s binary
validation signal “VL” to the recited “parameter” (see Final Act. 12; Ans.
49)—a signal that is routed to both histogram blocks 25 and 26 as shown in
the partial detail view of Pirim’s Figure 12 below:

Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

13

Partial detail view of Pirim’s Figure 12 showing binary validation signal
“VL” routed to histogram blocks 25 and 26

As Pirim’s page 23 explains, the binary validation signal “VL”
indicates whether the result of the speed and oriented direction is valid.
With this signal, the system can distinguish a valid output with the speed and
direction signals “V” and “DI” both being zero, where the signal is “1” for a
valid output, and “0” for an invalid output.
Despite Pirim’s routing the binary validation signal “VL” to both
histogram blocks 25 and 26 in Figure 12, it is not a parameter in the context
of the claimed invention, at least in the sense that the histograms that are
formed by Pirim’s histogram formation blocks 24 to 29 do not represent
such a parameter as a function of a validation signal—a key requirement of
claim 1. In this sense, the binary validation signal “VL” differs from the
other signals, namely: (1) luminance; (2) speed; (3) direction; (4) time
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

14
constant; (5) x position; and (6) y position that are represented by their
associated histograms from their respective histogram formation blocks.
That the earlier panel did not find that this particular validation signal
is a “parameter” as claimed, at least in the sense that it is treated by at least
two histogram calculation units to form a histogram representative of that
parameter as a function of a validation signal, only further underscores this
point. See ’336 IPR 46–47. That is, earlier panel found that the only signals
in Pirim’s Figure 12 that correspond to a recited “parameter” are those
directed solely to a respective histogram formation block 24 to 29, namely:
(1) luminance; (2) speed; (3) direction; (4) time constant; (5) x position; and
(6) y position. See id. The annotated version of Pirim’s Figure 12 on page
46 of the ’336 IPR decision is telling in this regard, for it labels each of the
six signals above clearly and unambiguously, yet does not label the binary
validation signal “VL” as shown below.

Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

15

Pirim’s annotated Figure 12 on page 46 of the ’336 IPR Board decision

This omission is consistent with the earlier panel’s finding that no two
histogram formation blocks are configured to process the same parameter as
required by claim 1. Id. at 47. It follows, then, that the binary validation
signal “VL” is not such a parameter despite its routing to histogram
formation blocks 25 and 26. Therefore, the Examiner’s mapping Pirim’s
binary validation signal “VL” to the recited “parameter” (Final Act. 12; Ans.
49) is inconsistent with the earlier panel’s findings in this regard and is,
therefore, untenable on this record.
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

16
Therefore, we are persuaded that the Examiner erred in rejecting
claim 1 as obvious over Pirim and Siegel. Because this issue is dispositive
regarding our reversing the Examiner’s rejection of this claim, we need not
address Appellant’s other associated arguments.

THE OBVIOUSNESS REJECTION OVER PIRIM AND HIROTA
Regarding independent claim 1, the Examiner reiterates the findings
from Pirim and its acknowledged deficiencies, but cites Hirota for teaching
using two “histogram calculation units,” namely memories 202 and 204 in
Figure 13, to process the same color parameter. Final Act. 16–18.
According to the Examiner, it would have been obvious to add another
undisclosed element, which the Examiner labels “204a,” that would be
adjacent and similar to element 204 in Hirota’s Figure 13 to process the
pixels’ video or luminance data.4 Final Act. 17; Ans. 56. Alternatively, the
Examiner concludes that it would have been obvious to use two or more
“copies” of Hirota’s system in Figure 13 to form two or more histograms of
each parameter. Final Act. 17. In yet another alternative, the Examiner
concludes that it would have been obvious to configure two of Pirim’s
existing histogram calculation units to treat the same parameter. Id.
On this record, we find these conclusions problematic essentially for
the reasons indicated by Appellant. First, it is unclear how or why an
additional undisclosed element similar to memory 204 in Hirota’s Figure 13,

4 In response to Appellant’s argument (App. Br. 57–61), the Examiner
clarifies that the additional undisclosed element would be added to Hirota’s
Figure 13. See Ans. 56; see also Reply Br. 29–32 (noting this clarification).
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

17
which the Examiner labels “204a,” would be added adjacent to that memory
in that figure to process the pixels’ video or luminance data as the Examiner
proposes. Final Act. 17; Ans. 56. As Appellant indicates (App. Br. 60–61;
Reply Br. 29–32), the Examiner’s proposal to somehow modify Hirota with
its own teachings to ostensibly arrive at the claimed invention by combining
this modification with Pirim is the product of impermissible hindsight and is,
therefore, untenable on this record.
Nor do we find availing the Examiner’s conclusion that it would have
been obvious to configure at least two of Pirim’s histogram calculation units
to treat one parameter as the Examiner proposes to increase “flexibility”
even assuming, without deciding, that Hirota’s memories 202 and 204 in
Figures 4 and 13 are “histogram calculation units” consistent with the
Examiner’s mapping. See Final Act. 16–18. As noted in the Abstract,
Hirota discloses a copying machine that determines automatically a
document type, such as color and black-and-white documents, based on
histograms. Hirota’s Figures 4 and 13, which differ only with respect to an
edge detection circuit 220 and AND gate 222 in Figure 13, show a histogram
generator whose memories 202 and 204 each receive a value signal “VH”
that is based on 8-bit R, G, and B input data. See Hirota, col. 7, ll. 20–37;
col. 16, l. 63–col. 17, l. 66. In both figures, a comparator 218 compares the
difference between the minimum and maximum R, G, B input data to a
reference value. Hirota, col. 8, ll. 15–22, col. 17, ll. 24–31. If the reference
value is lower in Hirota’s Figure 4, the data is written to memory 204.
Hirota, col. 8, ll. 15–22. But when the reference value is lower in Figure 13,
the data, along with the output from the edge detection circuit, is sent to the
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

18
AND gate whose output is sent to memory 204. See Hirota, col. 17, ll. 24–
45.
In either case, two separate components, namely the histogram
memories 202 and 204, each receive a color-based value signal “VH” and, in
that sense, treat the same color-based parameter. But to say that it would
have been obvious to somehow use two or more “copies” of Hirota’s system
in Figure 13 to form two or more histograms of each parameter in Pirim’s
system as the Examiner indicates (Final Act. 17) strains reasonable limits on
this record. To be sure, an obviousness determination based on teachings
from multiple references does not require an actual, physical substitution of
elements. In re Mouttet, 686 F.3d 1322, 1332 (Fed. Cir. 2012) (citations
omitted). Nor is the test for obviousness whether a secondary reference’s
features can be bodily incorporated into the structure of the primary
reference. In re Keller, 642 F.2d 413, 425 (CCPA 1981). Rather, the test is
what the combined teachings of the references would have suggested to
those of ordinary skill in the art. Id.
Nevertheless, the Examiner’s reasoning in this regard is unclear at
best. First, it is unclear on this record how “copies” of Hirota’s system in
Figure 13 could or would be used in Pirim’s system to treat the same
parameter which, as noted previously, is designed such that each histogram
formation block 24 to 29 treats a different parameter—a point emphasized
by the earlier panel in the ’336 IPR. See ’336 IPR 46–47. To the extent that
the Examiner concludes that it would been obvious to somehow modify one
or more of Pirim’s histogram formation blocks, such as block 24 that enables
forming a histogram for luminance values based on a received delayed
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

19
digital video signal “SR,” with some sort of dual-memory arrangement that
receives a color-based value signal such as that shown in Hirota’s Figures 4
or 13 (see Ans. 63), we fail to see how or why such an arrangement could or
would be used in connection with any of Pirim’s histogram formation blocks
24 to 29 in Figure 12 that process entirely different parameters, namely
luminance, speed, direction, time constant, and x and y positions,
respectively. In short, no two blocks are configured to process the same
parameter, let alone a color parameter. See Pirim 25–26; see also ’336 IPR
47 (noting these different parameters). To the extent that the Examiner
concludes that modifying one or more of Pirim’s histogram formation blocks
24 to 29 to somehow treat a color parameter that is not even contemplated
by Pirim, either by (1) including a dual-memory arrangement such as that in
Hirota in one or more of Pirim’s histogram formation blocks; (2) somehow
modifying two or more those blocks to treat the same parameter; or even (3)
adding another such block to the existing six in Pirim (see Ans. 60–61),
there is no persuasive evidence on this record to substantiate such a
conclusion.
The Examiner’s “increased flexibility” rationale in this regard is, at
best, speculative, even leaving aside the Examiner’s somewhat dubious
finding that Pirim’s image processing system is “generic” (Final Act. 17–18)
despite its sole drowsiness detection function as Appellant indicates. See
App. Br. 22–29; Reply Br. 4–5. Accord Bovik Decl. ¶ 28 (noting that Pirim
modifies a generic image processing system to detect drowsiness). That is,
the notion that that the proposed combination would ostensibly enable
Pirim’s “generic” image processing system to be more powerful in other
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

20
applications by detecting whether a piece of paper is color or black and
white as the Examiner indicates (Final Act. 18; Ans. 62) not only is
speculative and unsubstantiated, but runs counter to Pirim’s sole purpose—
detecting drowsiness. To the extent that the Examiner concludes that, in
view of Hirota’s copying machine functionality, it would have been obvious
for Pirim’s system to detect a particular type of document automatically
based on histograms in addition to, or in lieu of, detecting drowsiness—an
entirely different function—such a conclusion is, at best, speculative and
strains reasonable limits on this record. Accord Bovik Decl. ¶¶ 25–35
(noting this point).
To be sure, familiar items may have obvious uses beyond their
primary purposes, and often ordinarily skilled artisans can fit multiple
references’ teachings together like puzzle pieces. See KSR Int’l Co. v.
Teleflex, Inc., 550 U.S. 398, 420 (2007). But even assuming, without
deciding, that Pirim’s system could be used beyond its primary drowsiness
detection purpose, to suggest that it would be used to detect document types
automatically, as in Hirota’s copying machine under the Examiner’s
proposed combination, such a conclusion has not been substantiated on this
record apart from mere speculation. In short, the Examiner’s proposed
combination of the cited references is not supported by articulated reasoning
with rational underpinning to justify the Examiner’s obviousness conclusion.
Therefore, we are persuaded that the Examiner erred in rejecting
claim 1 as obvious over Pirim and Hirota. Because this issue is dispositive
regarding our reversing the Examiner’s rejection of this claim, we need not
address Appellant’s other associated arguments.
Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

21

CONCLUSION
The Examiner erred in rejecting claim 1 under 35 U.S.C. § 103.

DECISION
The Examiner’s decision to reject claim 1 is reversed.
Requests for extensions of time in this ex parte reexamination
proceeding are governed by 37 C.F.R. § 1.550(c). See 37 C.F.R. § 41.50(f).

REVERSED

rvb

Appeal 2019-004305
Reexamination Control 90/014,056
Patent US 6,959,293 B2

22

PATENT OWNER:

BYRNE POH LLP / IPT
11 BROADWAY
SUITE 760
NEW YORK, NY 10004

THIRD PARTY REQUESTER:

MARC PENSABENE
O'MELVENY & MYERS LLP
7 TIMES SQUARE,
NEW YORK, NY 10036