DivX LLC

Patent Trials and Appeals Board

Apr 22, 2021

IPR2020-00052 (P.T.A.B. Apr. 22, 2021)

Trials@uspto.gov Paper 82
571-272-7822 Date: April 22, 2021

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

NETFLIX INC. and HULU, LLC,
Petitioner,

v.

DIVX, LLC,
Patent Owner.
____________

IPR2020-00052
Patent 8,139,651 B2
____________

Before MONICA S. ULLAGADDI, SCOTT B. HOWARD, and
IFTIKHAR AHMED, Administrative Patent Judges.

AHMED, Administrative Patent Judge.

JUDGMENT
Final Written Decision
Determining No Challenged Claims Unpatentable
Granting In Part Petitioner’s Motion to Exclude
Granting In Part Patent Owner’s Motion to Exclude
35 U.S.C. § 318(a)

IPR2020-00052
Patent 8,139,651 B2

2
I. INTRODUCTION
This is a Final Written Decision in an inter partes review challenging
the patentability claims 1, 2, 4, and 17–19 (the “challenged claims”) of U.S.
Patent No. 8,139,651 B2 (Ex. 1001, “the ’651 patent”). We have
jurisdiction under 35 U.S.C. § 6.
Petitioner has the burden of proving unpatentability of the challenged
claims by a preponderance of the evidence. 35 U.S.C. § 316(e). Having
reviewed the parties’ arguments and supporting evidence, for the reasons
discussed below, we find that Petitioner has not demonstrated by a
preponderance of the evidence that the challenged claims are unpatentable.
Additionally, for the reasons discussed below, we grant in part, deny in part,
and dismiss in part the parties’ motions to exclude evidence (Papers 72, 73).
II. BACKGROUND
A. Procedural History
Netflix Inc. and Hulu, LLC (collectively “Petitioner”) requested an
inter partes review of the challenged claims of the ’651 patent. Paper 1
(“Petition” or “Pet.”). In view of the preliminary record, we concluded that
Petitioner satisfied the burden, under 35 U.S.C. § 314(a), to show that there
was a reasonable likelihood that Petitioner would prevail with respect to at
least one of the challenged claims. Accordingly, we instituted an inter
partes review of all the challenged claims on the asserted grounds. Paper 31
(“Inst. Dec.”); Paper 42 (public version).
After institution, DivX, LLC (“Patent Owner”) filed a Response.
Paper 54 (“PO Resp.”). Petitioner filed a Corrected Reply. Paper 671 (“Pet.

1 Petitioner’s original Reply (Paper 64) was expunged on request of the
parties.
IPR2020-00052
Patent 8,139,651 B2

3
Reply”). Patent Owner filed a Sur-reply. Paper 71 (“Sur-reply”). On
February 1, 2021, we held an oral hearing, the transcript of which is in the
record. Paper 81 (“Tr.”).
B. Related Proceedings
The ’651 patent is asserted in DivX, LLC v. Netflix, Inc., No. 2:19-cv-
01602 (C.D. Cal.), and DivX, LLC v. Hulu, LLC, No. 2:19-cv-01606 (C.D.
Cal.). Pet. 87; Paper 5, 1. Unified Patents Inc. filed an inter partes review
petition relating to the ’651 patent on which the Board entered judgment on
February 8, 2021. IPR2019-01379 (“Unified IPR”), Paper 52.
C. The ’651 Patent
The ’651 patent, titled “Video Deblocking Filter,” was filed on
May 26, 2010, and claims priority to a provisional application filed on
September 20, 2004. Ex. 1001, codes (54), (22), (60).
The ’651 patent concerns a method for “deblocking” a reconstructed
video frame. Id. at code (57). “Digital video sequences are composed of
frames of pixels, where the characteristics of the pixels are represented using
digital information.” Id. at 1:17–19. “Encoding schemes, such as the
scheme described in the MPEG-4 standard, can include video compression
algorithms that divide frames into blocks of pixels and use the characteristics
of the pixels within the blocks to encode the blocks of the video frame,”2
resulting in “artifacts at block boundaries when an encoded video frame is
reconstructed.” Id. at 1:25–31. Those artifacts can be removed from a
reconstructed image by “applying a deblocking filter to pixels adjacent block
boundaries.” Id. at 1:32–34.

2 The MPEG-4 standard was developed by the Motion Picture Experts
Group. Id. at 1:35–36.
IPR2020-00052
Patent 8,139,651 B2

4
Figure 1 of the ’651 patent is reproduced below.

Figure 1, above, illustrates a portion of reconstructed video frame 10, made
up of a number of pixels 12. Id. at 8:6–8. Each block is made up of an 8 x 8
area of pixels, and the deblocking filter can be applied to the pixels both
along horizontal boundaries 14 and vertical boundaries 16 between the
reconstructed blocks. Id.
Figure 2 of the ’651 patent is reproduced below.

IPR2020-00052
Patent 8,139,651 B2

5
Figure 2, above, illustrates the steps for applying a deblocking filter by
“identifying (22) the location[s] of horizontal and vertical block
boundaries,” determining (24) “[t]he level of detail of the region of the video
frame in which the block boundary is located,” and applying either dc offset
filter (28) or another filter appropriate to level of detail (30) to pixels
adjacent the boundary depending on a determination (26) of whether the
boundary lies in a smooth region or in a region with a higher level of detail.
Id. at 8:20–31. A deblocking filter can be applied to pixel chrominance
and/or to pixel luminance. Id. at 8:34–37.
D. Challenged Claims
Petitioner challenges claims 1, 2, 4, and 17–19. Challenged claim 1
is the only independent claim and is reproduced below.
1. A method of deblocking a reconstructed video frame,
comprising:
identifying a boundary between two blocks of the
reconstructed video frame;
determining the level of detail of the reconstructed video
frame across a region in which the block boundary is located,
wherein the region includes pixels from multiple rows and
multiple columns of the reconstructed video frame that
encompass pixels immediately adjacent to at least two sides of
the block boundary and includes at least one pixel that is not
immediately adjacent to the block boundary;
selecting a filter to apply to predetermined pixels on
either side of the block boundary based upon the determined
level of detail.

Ex. 1001, 13:7–22.
IPR2020-00052
Patent 8,139,651 B2

6
E. Prior Art and Instituted Grounds of Unpatentability
We instituted trial based on the following ground of unpatentability:
Claims Challenged 35 U.S.C. §3 Reference(s)/Basis
1, 17, 18 102(b) Vehviläinen4
1, 17–19 103(a) Vehviläinen
1, 2, 4, 17–19 103(a) Vehviläinen, Kadono5
Inst. Dec. 57–58; Pet. 9, 19–86.
In support of its patentability challenges, Petitioner relies on the
declaration testimony of Dr. Victor Michael Bove, Jr. See Ex. 1003. Patent
Owner relies on the declaration testimony of Dr. Chandrajit Bajaj. See
Ex. 2050.
III. ANALYSIS
A. Principles of Law
“In an [inter partes review], the petitioner has the burden from the
onset to show with particularity why the patent it challenges is
unpatentable.” Harmonic Inc. v. Avid Tech., Inc., 815 F.3d 1356, 1363 (Fed.
Cir. 2016) (citing 35 U.S.C. § 312(a)(3) (requiring inter partes review
petitions to identify “with particularity . . . the evidence that supports the
grounds for the challenge to each claim”)). This burden of persuasion never

3 Because the application leading to the ’651 patent was filed before
March 16, 2013, patentability is governed by the versions of 35 U.S.C.
§§ 102 and 103 preceding the Leahy-Smith America Invents Act (“AIA”),
Pub L. No. 112–29, 125 Stat. 284 (2011).
4 U.S. Patent No. 6,504,873 B1, issued Jan. 7, 2003 (Ex. 1004,
“Vehviläinen”).
5 U.S. Patent Appl. Pub. No. 2004/0076237 A1, published Apr. 22, 2004
(Ex. 1005, “Kadono”).
IPR2020-00052
Patent 8,139,651 B2

7
shifts to Patent Owner. Dynamic Drinkware, LLC v. Nat’l Graphics, Inc.,
800 F.3d 1375, 1378 (Fed. Cir. 2015) (discussing the burden of proof in
inter partes review).
To establish anticipation, each and every element in a claim, arranged
as recited in the claim, must be found in a single prior art reference.
Net MoneyIN, Inc. v. VeriSign, Inc., 545 F.3d 1359, 1369 (Fed. Cir. 2008);
Karsten Mfg. Corp. v. Cleveland Golf Co., 242 F.3d 1376, 1383 (Fed. Cir.
2001). Although the elements must be arranged or combined in the same
way as in the claim, “the reference need not satisfy an ipsissimis verbis test,”
i.e., identity of terminology is not required. In re Gleave, 560 F.3d 1331,
1334 (Fed. Cir. 2009); accord In re Bond, 910 F.2d 831, 832 (Fed. Cir.
1990). Further, to be anticipating, a prior art reference must be enabling and
must describe the claimed invention sufficiently to have placed it in
possession of a person of ordinary skill in the field of the invention. Helifix
Ltd. v. Blok-Lok, Ltd., 208 F.3d 1339, 1346 (Fed. Cir. 2000); In re Paulsen,
30 F.3d 1475, 1479 (Fed. Cir. 1994).
As set forth in 35 U.S.C. § 103(a),
[a] patent may not be obtained . . . if the differences between the
subject matter sought to be patented and the prior art are such
that the subject matter as a whole would have been obvious at the
time the invention was made to a person having ordinary skill in
the art to which said subject matter pertains.
The question of obviousness is resolved on the basis of underlying factual
determinations including: (1) the scope and content of the prior art; (2) any
differences between the claimed subject matter and the prior art; (3) the level
of ordinary skill in the art; and (4) when in evidence, objective evidence of
nonobviousness. Graham v. John Deere Co., 383 U.S. 1, 17–18 (1966).
IPR2020-00052
Patent 8,139,651 B2

8
An obviousness determination requires finding “a motivation to
combine accompanied by a reasonable expectation of achieving what is
claimed in the patent-at-issue.” Intelligent Bio-Sys., Inc. v. Illumina
Cambridge Ltd., 821 F.3d 1359, 1367 (Fed. Cir. 2016). Petitioner cannot
satisfy its burden of proving obviousness by employing “mere conclusory
statements.” In re Magnum Oil Tools Int’l, Ltd., 829 F.3d 1364, 1380 (Fed.
Cir. 2016). Instead, Petitioner must articulate a reason why a person of
ordinary skill in the art would have combined the prior art references. In re
NuVasive, 842 F.3d 1376, 1382 (Fed. Cir. 2016).
To prevail in its challenges to Patent Owner’s claims, Petitioner must
demonstrate by a preponderance of the evidence that the claims are
unpatentable. 35 U.S.C. § 316(e); 37 C.F.R. § 42.1(d) (2019).
Petitioner asserts that Vehviläinen anticipates certain claims of the
’651 patent, and that Vehviläinen, either alone or in combination with
Kadono, would have rendered the subject matter of certain claims of the
’651 patent obvious to one of ordinary skill in the art at the time of the
invention. Pet. 19–86. We analyze the asserted grounds of unpatentability
in accordance with these principles to determine whether Petitioner has met
its burden in establishing unpatentability of the challenged claims at trial.
B. Level of Ordinary Skill in the Art
We review Petitioner’s asserted obviousness grounds in view of the
understanding of a person of ordinary skill in the art at the time of the
invention. Graham, 383 U.S. at 17. Petitioner proposes that a person of
ordinary skill in the art “would have had a bachelor’s degree in electrical
engineering, computer engineering, computer science, or a related subject,
and at least two to three years of work experience in image and/or video
IPR2020-00052
Patent 8,139,651 B2

9
processing, including previous work on the development and application of
video compression and decompression methods, in particular deblocking
techniques.” Pet. 13 (citing Ex. 1003 ¶¶ 59–60). In our institution decision,
we adopted Petitioner’s proposed level without the qualifier “at least” as to
the years of work experience. Inst. Dec. 19–20.
Patent Owner argues that “[t]he proper level is that of a Bachelor’s
degree holder with 2–3 years of experience in video compression—not
experience with deblocking techniques in particular.” PO Resp. 12.
According to Patent Owner, deblocking is merely a specialized technique in
the video compression field, not a separate field of its own. Id. at 12–13
(citing Ex. 1001, 1:15–16; Ex. 2050 ¶¶ 26–28). Patent Owner contends that
while video compression is a widely recognized field, it would not have
been ordinary for persons of this education and experience level in video
compression to have had experience specifically in deblocking compressed
video and related techniques. Id. at 13–14 (citing Ex. 2081; Ex. 2046; Ex.
2047; Ex. 2052; Ex. 2053; Ex. 2050 ¶¶ 29–31).
Petitioner responds that a person of ordinary skill in the art
(“POSITA”) is presumed to be aware of all pertinent prior art, that Dr. Bajaj
admitted that a person of ordinary skill in the art “would be knowledgeable
about the blocking methods . . . because it was also adopted by the MPEG-4
standard at that time,” and that “deblocking is necessarily part of the
decoding.” Pet. Reply 17 (citing Ex. 1017, 45:19–46:2, 51:19–52:14,
44:19–25, 46:9–47:5, 151:10–152:25, 144:20–145:18). Petitioner therefore
contends that deblocking is not a niche field.
Patent Owner responds that even assuming Petitioner’s assertions are
correct, that still does not address the point that a person of ordinary skill in
IPR2020-00052
Patent 8,139,651 B2

10
the art would not have been ordinarily able to develop a new deblocking
scheme. Sur-reply 19.
We are not persuaded that the level of ordinary skill in the art requires
previous work experience in development of deblocking techniques. The
person of ordinary skill in the art is a hypothetical person who is presumed
to have known the relevant art at the time of the invention. In re GPAC,
Inc., 57 F.3d 1573, 1579 (Fed. Cir. 1995). Factors that may be considered in
determining the level of ordinary skill in the art include, but are not limited
to, the types of problems encountered in the art, the sophistication of the
technology, and educational level of active workers in the field. Id. In a
given case, one or more factors may predominate. Id.
The ’651 patent relates generally to video compression. Ex. 1001,
1:15. Additionally, the prior art relied upon by the examiner during
prosecution of the ’651 patent also relates generally to video compression,
not just to deblocking techniques. See Ex. 1002, 73–81; see also Ex. 1010,
codes (54), (57) (the Kim reference cited by the examiner, titled “Reducing
Blocking and Ringing Artifacts in Low Bit-Rate [Video] Coding”).
Although we agree with the parties that deblocking would have been part of
decoding techniques related to video compression and that a person of
ordinary skill in the art would have been knowledgeable of deblocking, we
are not persuaded that work experience in those specific techniques is
necessary.6 As long as a person has had work experience in the
development and application of video compression and decompression

6 Notably, Petitioner discusses other related concepts of video processing
that a POSITA would have been familiar with but does not include those in
its proposed definition of the level of ordinary skill. Pet. 13–14.
IPR2020-00052
Patent 8,139,651 B2

11
methods and Petitioner’s proposed educational background in an electrical/
computer engineering or science discipline, that would sufficiently prepare a
person for work in the art pertaining to the ’651 patent. Accordingly, the
level of ordinary skill in the art of the ’651 patent is that of a person having a
bachelor’s degree in electrical engineering, computer engineering, computer
science, or a related subject, and two to three years of work experience in
image and/or video processing, including previous work on the development
and application of video compression and decompression methods. Our
determination is consistent with the Board’s determination of the level of
ordinary skill for the ’651 patent in the Unified IPR. See IPR2019-01379,
Paper 52, 9–10.
We note, though, that we do not find any instance in which the
differences between the parties’ respective descriptions of the level of
ordinary skill in the art impacts our findings or conclusions in this case. See
infra §§ III.E–III.F. The parties also do not identify any specific dispute that
is impacted by the differences between the parties’ respective descriptions.
See, e.g., PO Resp. 14 (arguing that “the Petition fails regardless of which
level of ordinary skill is adopted by the Board”).
C. Claim Construction
We apply the same claim construction standard that would be used in
a civil action under 35 U.S.C. § 282(b), following the standard articulated in
Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc). 37 C.F.R.
§ 42.100(b) (2019). In applying this standard, we generally give claim terms
their ordinary and customary meaning, as would be understood by a person
of ordinary skill in the art, at the time of the invention and in the context of
the entire patent disclosure. Phillips, 415 F.3d at 1312–13.
IPR2020-00052
Patent 8,139,651 B2

12
In our institution decision, we construed the term “level of detail” as
“level of variation in visual elements across adjacent pixels.” Inst. Dec. 20–
22. Patent Owner asserts that the Board’s construction is correct, supported
by the specification and the testimony of both parties’ experts. PO Resp.
15–16 (citing Ex. 1001, 1:52–57, Ex. 2050 ¶¶ 53–57; Ex. 2049, 69:6–12).
Patent Owner contends that the ’651 patent “expressly defines the claimed
‘detail’ in this context: The term ‘detail’ is typically used to describe visual
elements of a video frame that vary significantly across adjacent pixels.” Id.
at 15 (quoting Ex. 1001, 1:55–57).
Petitioner does not disagree with our preliminary construction. Pet.
Reply 5 (“This is consistent with the Petition and Dr. Bove’s analysis.”).
Petitioner, however, argues that the manner in which Patent Owner applies
the construction improperly narrows the claims to the sum of absolute
differences (“SAD”) formula disclosed in the specification and claimed. Id.
at 7–8.
For the reasons discussed in our institution decision (Inst. Dec. 20–
22), which we incorporate by reference, we maintain our prior construction
of the term “level of detail” as “level of variation in visual elements across
adjacent pixels.” As the parties point out, the Specification defines the term
“detail” consistent with its ordinary meaning as the amount of variation in
visual elements of a video frame across adjacent pixels. Ex. 1001, 1:55–57;
PO Resp. 15; Pet 16. The Specification provides examples of calculating the
level of detail using chrominance or luminance values. See, e.g., id. at 2:5–
7, 3:56–57, 4:1–2. The Specification further supports that the calculation of
the variation is done for adjacent pixels, and does not use the term in a
manner inconsistent with the typical understanding. See e.g., id. at 3:34–36,
IPR2020-00052
Patent 8,139,651 B2

13
3:41–42, 4:32–33. We therefore maintain that the meaning of “level of
detail” is “level of variation in visual elements across adjacent pixels.”
We address Petitioner’s arguments regarding Patent Owner’s
application of the construction to the prior art as part of our obviousness
analysis. See infra § III.E.1.a.4.
D. Overview of the Asserted Prior Art
1. Vehviläinen (Ex. 1004)
Vehviläinen relates to a method of filtering compressed video “based
on the amount of activity inside the adjacent video blocks and the activity at
the boundary between the adjacent video blocks.” Ex. 1004, codes (54),
(57). Vehviläinen discloses that
If there is [a] lot of activity inside two adjacent video blocks, and
at the boundary between them, that part of video frame most
probably includes important information, and filtering at that
area should be outlined to the minimum or completely omitted.
On the other hand, if there is no or very little (probably just noise)
activity in that area, the area is supposed to be “smooth” (e.g., a
wall of one color), and an efficient filter can be used for a
relatively large area to minimize the visible blocking effect,
which is as its worst just at the “smooth” areas.
Id. at 2:46–55.
IPR2020-00052
Patent 8,139,651 B2

14
Figure 5A of Vehviläinen is reproduced below.

Figure 5A, above, illustrates a pixel area that is used for calculation of
picture activity within an area of the video frame. Id. at 4:11–14.
Vehviläinen discloses that calculation of variance inside the block N is done
using the 12 video pixels (51) marked with black circles, using equation (4),
which follows:

Equation (4) uses the difference between the largest and smallest pixel value
within a given set of pixels. Id. at 10:44–55. It is an estimation of
Vehviläinen’s variance calculation method provided in equation (3), which
follows:
1
1 ̅
Vehviläinen also discloses a selection process for the filtering method
to be used based on whether a given area is smooth, moderate, or high
activity.
IPR2020-00052
Patent 8,139,651 B2

15
Figure 6 of Vehviläinen is reproduced below.

Figure 6, above, illustrates a flowchart for the filter selection process of
Vehviläinen, including conditional steps 66, 70, and 74 that determine the
filter to be applied in steps 80, 82, and 84. Id. at 11:4–12:6. Vehviläinen
uses the same conditional equation for each of conditional steps 66, 70, and
74 to determine whether filtering is required at the boundary between
adjacent video blocks. Id. at 9:51–53, 10:5–9. This conditional equation,
labeled “equation (2),” is reproduced below:
IPR2020-00052
Patent 8,139,651 B2

16

In equation (2) above, “activityIN is a variance inside the block, activityEDGE
is a variance over the edge between adjacent blocks, THRESHOLDIN is a
threshold value, [and] MQUANT is a quantization value for the block under
processing.” Id. at 9:51–66, 10:6–9.
2. Kadono (Ex. 1005)
Kadono discloses methods related to selectively applying a
deblocking filter to compressed video by first determining whether
deblocking is necessary. Ex. 1005, code (57). In one embodiment, Kadono
discloses comparing pixel values for groups of pixels that are symmetrically
located on left and right sides of the block boundary. Id. ¶ 204. Figure 19 of
Kadono is reproduced below.

Figure 19, above, illustrates groups of four pixels r1 and r2 that Kadono’s
process compares to determine the level of deblocking to be applied to the
pixels on either side of the block boundary. Id.
IPR2020-00052
Patent 8,139,651 B2

17
Figure 6 of Kadono is reproduced below.

Figure 6, above, is a flowchart showing how the level of filtering to be
applied to pixels is determined based on a comparison of the difference
between pixel values on either side of a block boundary to a threshold value,
Ω, according to one of Kadono’s embodiments. Id. ¶¶ 120–129, Figs. 6,
8(b).
E. Anticipation and Obviousness based on Vehviläinen
Petitioner contends that claims 1, 17, and 18 are unpatentable as
anticipated by or obvious over Vehviläinen. Pet. 19–49. Petitioner also
contends that claim 19 is unpatentable as obvious over Vehviläinen. Id. at
49–52. For the reasons that follow, we are not persuaded that Petitioner has
established by a preponderance of the evidence that these claims are
unpatentable under §§ 102 or 103(a) in view of Vehviläinen.
IPR2020-00052
Patent 8,139,651 B2

18
1. Independent Claim 1
Patent Owner argues that Vehviläinen fails to teach or suggest the
“determining the level of detail” limitation of independent claim 1. PO
Resp. 14–22. We, therefore, begin our discussion with the parties’
arguments on this limitation.
a) “determining the level of detail of the reconstructed
video frame across a region in which the block boundary
is located, wherein the region includes pixels from
multiple rows and multiple columns of the reconstructed
video frame that encompass pixels immediately adjacent
to at least two sides of the block boundary and includes
at least one pixel that is not immediately adjacent to the
block boundary;”
(1) The Petition
Petitioner contends that Vehviläinen’s calculation of activityIN as a
measure of variance for the pixels marked by dots in Figure 5A teaches
determining the level of detail in a region of the reconstructed video frame
where the block boundary is located. Pet. 33 (citing Ex. 1004, 9:50–67,
11:30–41, Fig. 5A; Ex. 1003 ¶ 113).7 Petitioner reproduces an annotated
version of Vehviläinen’s Figure 5A (id. at 34), as shown below.

7 Petitioner presents the same contentions with respect to this limitation
under both its anticipation and obviousness grounds based on Vehviläinen.
Pet. 33–43. We therefore address these grounds together.
IPR2020-00052
Patent 8,139,651 B2

19

Petitioner’s annotated Figure 5A of Vehviläinen, above, show an 8 x 3 pixel
area 54, comprising two regions 51 and 52 on either side of a vertical block
boundary, annotated in red, and a third 2 x 3 region 53 comprising pixels
immediately adjacent to the block boundary. Pet. 33–34 (citing Ex. 1004,
10:1–15, 11:29–41, Fig. 5A; Ex. 1003 ¶ 114). Referring to Figure 5A,
Petitioner argues that Vehviläinen analyzes the pixels marked by dots to
determine if the block boundary lies in a “smooth area.” Id. According to
Petitioner, Vehviläinen’s teaching “is mirrored by embodiments of the ’651
patent, which similarly analyze whether the block boundary is ‘located
within very smooth regions of a video frame’ or whether it ‘lies in a region
possessing a high level of detail,’” and “calculates the level of detail based
on the level of variation across pixels in the region.” Id. (citing Ex. 1001,
1:45–63).
Petitioner contends that Vehviläinen’s “min-max approximation” of
pixel variance “takes the difference between the largest and smallest pixel
values in an area.” Pet. 35 (citing Ex. 1004, 9:51–67, 10:44–55; Ex. 1003
¶ 115). Petitioner further contends that Vehviläinen teaches
IPR2020-00052
Patent 8,139,651 B2

20
calculating the variance of pixels on the left side of the block
boundary, shown in Figure 5A as marked pixels (dots) in block
N, labelled 51, then calculating the variance of pixels on the right
side of the block boundary, shown as marked pixels in block
N+1, labelled 52, and then averaging the two variance
calculations into a single variance value (‘activityIN’) for the
area.
Id. (citing Ex. 1004, 11:29–46, 11:56–12:10, Fig. 5A; Ex. 1003 ¶ 115); see
also id. (discussing also the numerical examples provided by Vehviläinen
for each of those calculations). Petitioner further asserts that “[t]he variance
at the boundary is called ‘activityEDGE’ and ‘is calculated using formula (5)’
with the 6 pixels labelled (53).” Id. at 35–36. According to Petitioner,
Vehviläinen teaches equation (2), which uses those variance parameters,
activityIN and activityEDGE, together with THRESHOLDIN and MQUANT
values to determine whether to apply a smooth filter. Id. at 36. Petitioner
asserts that calculations for the moderate and high activity areas are
performed “in a similar manner.” Id. at 36 (citing Ex. 1004, 11:56–12:10,
Figs. 5B–5C; Ex. 1003 ¶ 115). Petitioner thus argues that Vehviläinen
teaches determining a level of detail (activityIN) across a region (the dotted
area in Figure 5A) that includes a block boundary (the center line in Figure
5A) by taking an average of the variance calculations in the region. Id.
(citing Ex. 1004, 11:29–46; Ex. 1003 ¶ 116).
Petitioner further argues that a person of ordinary skill in the art
would have found it obvious to apply Vehviläinen’s teaching of taking the
average of the variance calculations for determining the level of detail of the
region containing all marked pixels in Figure 5A. Id. (citing Ex. 1003
¶ 116). Petitioner contends that doing so would have provided “a simple,
efficient, and accurate method for determining the level of detail of the
IPR2020-00052
Patent 8,139,651 B2

21
region.” Id. Petitioner contends that it would have also been obvious for
one of ordinary skill in the art to determine the level of detail for just the
“smooth” area (the one shown in Figure 5A) and then select the smooth
filter, if the area is smooth, or the high/moderate activity filter otherwise.
Pet. 38, 39 (citing Ex. 1003 ¶ 120).
(2) Patent Owner’s Response
In response, Patent Owner argues that the Petition relies on an
incorrect understanding of the claimed “level of detail,” and that
“Vehviläinen’s variance calculation (or its approximation proxy) does not
determine the level of change across adjacent pixels, as required under the
proper construction of ‘level of detail.’” PO Resp. 16. As discussed above,
Patent Owner contends that our preliminary construction of the term “is
correct, supported by the specification and the testimony of both parties’
experts.” Id. at 15.
Patent Owner contends that Vehviläinen teaches not filtering adjacent
pixels “in comparison with each other,” and discloses a process of choosing
to apply filters consistent with that teaching. Id. at 16–17. Patent Owner
asserts that Vehviläinen’s method “[r]educes the disappearance of the
resolution in the picture in the filtering process according to the invention
because the adjacent video pixels as such are not filtered in comparison with
each other but are adjusted in comparison to the reference line.” Id. at 17
(quoting Ex. 1004, 2:41–45). Patent Owner annotates Figure 4 of
Vehviläinen, reproduced below.
IPR2020-00052
Patent 8,139,651 B2

22

Figure 4, above, annotated by Patent Owner, shows the forwarding of the
numerical values of each individual video pixel (shown in blue) towards the
reference line (shown in red) at the boundary between the adjacent video
blocks, when using smooth filtering mode. Ex. 1004, 4:8–10; PO Resp. 17–
18 (citing Ex. 1004 Fig. 4; 8:29–62. Ex. 2050 ¶¶ 59–60). Patent Owner
contends that Vehviläinen’s filter is designed to “maintain numerical
differences between adjacent video pixels” in order to preserve “the details
in the picture,” and “deliberately sets out to avoid founding its filtering
scheme upon the differences in the values of adjacent pixels, instead moving
all pixels towards a reference line.” PO Resp. 18 (citing Ex. 1004, 8:63–9:6;
Ex. 2050 ¶ 61).
Patent Owner argues that Vehviläinen’s filter selection process,
matching its filter application process, also does not analyze adjacent pixels
in comparison with each other, but instead, utilizes a statistical variance
calculation to determine whether or not to apply a particular filter to a
particular region. PO Resp. 18 (citing Ex. 1004, 9:51–67; 10:32–44). Patent
Owner contends that consistent with a POSITA’s understanding,
IPR2020-00052
Patent 8,139,651 B2

23
Vehviläinen discloses that “variance” is a known statistical calculation
measuring the difference between the value of each pixel in the region and
the average value of all pixels in the region, and not the difference in value
between adjacent pixels. Id. (citing Ex. 1004, 10:32–43, 2:41–45). To
highlight that distinction, Patent Owner provides “a side-by-side comparison
of an exemplary equation for calculating the level of detail in the [’651
patent] and the equation Vehviläinen teaches for the calculation of
variance.” Id. at 18–19 (citing Ex. 1001, 3:50; Ex. 1004, 10:35 (equation 3);
Ex. 2050 ¶¶ 62–65).
The calculation shown in equation 4 of Vehviläinen, Patent Owner
contends, “is simply a proxy for its variance calculation, and therefore does
not provide an assessment of level of change across ‘adjacent’ pixels for the
same reason that variance does not.” Id. at 20 (citing Ex. 2050 ¶ 66). Patent
Owner argues that if Vehviläinen’s variance (equation 3) fails to capture the
level of change across adjacent pixels, so too does its proxy min-max
estimation (equation 4). Id. (citing Ex. 1004, 10:32–55; Ex. 2049, 149:4–9;
Ex. 2050 ¶ 67). Moreover, Patent Owner contends, the min-max estimation
is not related to the level of change across adjacent pixels and fails to
capture what happens between adjacent pixels that lie between the pixels
with the minimum and maximum values. Id. at 21 (citing Ex. 2050 ¶ 68).
(3) Petitioner’s Reply
In its Reply, Petitioner responds that its analysis is consistent with our
construction of the term and that Patent Owner interprets the claim limitation
to limit it to the disclosed embodiments. Pet. Reply 5–9. Petitioner
contends that Patent Owner rewrites the Board’s construction for “level of
detail” as the level of variation in visual elements between adjacent pixels.
IPR2020-00052
Patent 8,139,651 B2

24
Id. at 7 (citing PO Resp. 18, 19, 32). According to Petitioner, Patent
Owner’s interpretation “narrows the claims to SAD formulas of exemplary
embodiments, which is improper because those SAD formulas are recited
within unchallenged dependent claims.” Id. at 7–8 (citing GE Lighting Sols.,
LLC v. AgiLight, Inc., 750 F.3d 1304, 1309–10 (Fed. Cir. 2014)).
Petitioner argues that Vehviläinen calculates variance across the
adjacent pixels in each region using the min-max approximation which
analyzes all pixels in the region to determine the largest and smallest value.
Id. at 5 (citing Ex. 1004, 10:32–55; Ex. 2049, 185:10–188:14 (“[I]f you take
the group of pixels inside a region and you calculate the variables across that
group of pixels, that will give you a good estimate as to whether this is a
high or low activity.”), 143:13–17). Petitioner further contends
“Vehviläinen uses variance and its min-max approximation as statistical
measures of variation to characterize how much the adjacent pixels in
regions 5A-5C vary from each other—and thus whether regions 5A-5C are
smooth, moderate, or high activity.” Id. at 6 (citing Ex. 1004, 9:51–60,
10:32–55, 11:29–12:10, Figs. 5A–5C; Ex.1003 ¶¶ 100–101, 115–116).
Petitioner faults Patent Owner for conflating the step of determining level of
detail with the subsequent filtering process. Id. (citing Inst. Dec. 48).
Even under an interpretation requiring that the level of detail be
determined across adjacent pixels, Petitioner argues Vehviläinen’s Figure 5C
teaches calculating variance across adjacent pixels because each row in that
figure has one pair of adjacent pixels on each side of the block boundary. Id.
at 9–10 (citing Ex. 1004, 11:67–12:3, 11:29–46).
IPR2020-00052
Patent 8,139,651 B2

25
(4) Analysis
On consideration of the full record, we agree with Patent Owner that
Vehviläinen does not teach or suggest determining the “level of detail” as
we have construed it––as the level of variation in visual elements across
adjacent pixels. We are not persuaded by Petitioner’s arguments (Pet. Reply
7) that our construction of the term, requiring evaluation of adjacent pixels,
improperly narrows the claim term to the disclosed embodiments. Instead,
the construction is based on what the Specification describes as the typical
understanding, i.e., the ordinary meaning of the term “detail” (Ex. 1001,
1:55–57)—which both parties agree on—and which is reflected in every
disclosed embodiment. For example, our construction does not require the
level of detail determination to be made using the sum of absolute
differences formula or for each adjacent pair of pixels in each row or column
of a given region, as performed by some of the disclosed embodiments. See
id. at 3:41–42, 3:50, 4:32–33. We agree with Petitioner that those
limitations are recited in some of the dependent claims. See, e.g., id. at
13:25–27 (dependent claim 2 requiring that “the determination of the level
of detail . . . further comprises taking the sum of the absolute difference of at
least some of the pixels within a set of pixels surrounding the block
boundary”), 13:39–51 (dependent claim 5 reciting a formula for calculating
the following sum). Our construction therefore embodies the ordinary
meaning of the term without importing further limitations of the disclosed
embodiments. Petitioner’s proposal, on the other hand, eliminates the
requirement that detail be calculated across adjacent pixels, improperly
broadening the limitation to a determination of variation in visual elements
IPR2020-00052
Patent 8,139,651 B2

26
across the entire claimed region. Petitioner offers no support in the
Specification or elsewhere for such an interpretation.
In view of our construction, we are not persuaded that Vehviläinen’s
variance teaches or suggests determining the level of detail. Vehviläinen
uses “variance” to determine activity in a given region. Ex. 1004, 10:32–43.
Vehviläinen explains,
Normal variance calculation is an exhausting operation, and
performed according to equation 3.
1
1 ̅
where
xi is the numerical picture value of each individual
video pixel, and
x is the mean value of all N samples.
Id. (emphasis added). Dr. Bajaj testifies, consistent with Vehviläinen’s
disclosed equation, that a “POSITA would understand, ‘variance’ is a
statistical calculation measuring the difference between the value of the
pixels in the region and the average value of all pixels in the region—not the
difference in value between adjacent pixels.” Ex. 2050 ¶ 63. We find that
testimony persuasive. Thus, equation 3 does not disclose determining the
“level of variation in visual elements across adjacent pixels.”
With regard to its equation (4), Vehviläinen clearly states that
equation (4) is merely an approximation of equation (3):
IPR2020-00052
Patent 8,139,651 B2

27
The variance can be estimated by a more simple method called
min-max approximation. This method defines the variance as

where
xmax is the largest one of numerical picture values of the
video pixels inside the picture area under
examination, and
xmin is the smallest one.
Ex. 1004, 10:44–55. Dr. Bajaj testifies that this calculation “is simply a
proxy for its variance calculation” and does not determine the level of
change across adjacent pixels. Ex. 2050 ¶¶ 65, 66. Dr. Bove does not
disagree with that assessment.8 Ex. 2049, 149:4–9 (“Q. So if it’s fair to say
that a person of ordinary skill in the art could understand that equation (4)
would provide a reasonable and acceptable estimation of equation (3) with
much less calculation? A. Yes, one would understand that.”). We agree that
equation 4 is a simpler method of arriving at the same variance calculation
as equation 3, and therefore, does not teach or suggest comparing adjacent
pixels to calculate activity within a region.
Even considering the min-max calculation as applied specifically to
high-activity regions (shown in Vehviläinen’s Figure 5C, reproduced
below), we do not agree with Petitioner that because each row has one pair

8 Patent Owner argues that we should find all of Dr. Bove’s testimony not
credible because he offered “uncandid testimony with respect to his career
history and qualifications.” PO Resp. 9–11 (citing Ex. 1003, 5, 123–125;
Ex. 2049, 173:11–20, 174:13–19; Ex. 2078; Ex. 2079). Petitioner responds
that Dr. Bove was truthful in his testimony. See Pet. Reply 27. We have
reviewed Patent Owner’s cited evidence and do not find any inconsistencies
in Dr. Bove’s testimony. We therefore find Dr. Bove’s testimony with
respect to his career history and qualifications credible.

IPR2020-00052
Patent 8,139,651 B2

28
of adjacent pixels on each side of the block boundary, calculating min-max
necessarily includes determining the level of variation in visual elements
across each pair of adjacent pixels. Pet. Reply 10.

Figure 5C, above, illustrates a pixel area that is used for calculation of
picture activity within a high activity area of the video frame. Ex. 1004,
4:11–14. The calculation of variance is performed for block 57 (similar to
block 52 of Figure 5A), applying equation (4) to the smallest and the largest
values of the six pixels of that block. Id. at 11:33–39. Although this
calculation takes into account each of the pixel values of the region, the
result of that determination does not reflect variation between adjacent
pixels and is instead driven by the difference between pixel values across the
entire region being evaluated.9

9 We agree with Patent Owner that the possibility alone that two adjacent
pixels may have a minimum and maximum pixel value and may
consequently be compared when evaluated by equation (4) is insufficient to
meet the claim limitation. See In re Facebook, Inc., 743 F. App’x. 998,
1001-1002 (Fed Cir. 2018) (reversing the Board’s anticipation and
obviousness determinations and deciding that an example in the prior art
reference showing that it is possible for image elements to be contiguous is
not sufficient to teach a claim limitation that requires contiguity); In re
IPR2020-00052
Patent 8,139,651 B2

29
Dr. Bajaj testifies that this method of calculating variance is key to
Vehviläinen’s filter design which seeks to maintain differences between
adjacent video pixels so as to preserve picture detail, and therefore moves
the values of adjacent pixels in unison towards a reference line. Ex. 2050
¶¶ 61, 62. This testimony is consistent with Vehviläinen’s disclosure:
The filtering method according to the invention is designed to
maintain numerical differences between adjacent video pixels in
each block, like the differences in numerical values of pixels 41,
42, 43 and 44 in block 5 and correspondingly the differences
between pixels 45, 46, 47 and 48. When the differences in the
numerical values of the video pixels are preserved, the details in
the picture will as well, and the resolution of the video picture
will preferably not decrease and small details will not vanish in
a disturbingly manner.
Ex. 1004, 8:63–9:6 (referring to Fig. 4, reproduced above); see also id. at
2:41–45. Dr. Bajaj further testifies that “Vehviläinen deliberately sets out to
avoid founding its filtering scheme upon the differences in the values of
adjacent pixels” and instead utilizes a variance calculation that looks to
difference between pixel values across an entire region to be filtered. Id.
(citing Ex. 1004, 9:51–67; 10:32–44). We credit Dr. Bajaj’s testimony
because it is consistent with Vehviläinen’s own discussion of its determining
and filtering step, and we are persuaded that Vehviläinen’s variance
calculation works in lock-step with its goal of preserving even small details.
We are therefore not persuaded by Petitioner’s argument and Dr. Bove’s
testimony that Vehviläinen’s min-max approximation evaluates “level of

Felton, 484 F.2d 495, 500 (CCPA 1973) (reversing anticipation and
obviousness determinations where “it would be mere happenstance if any
structure made according to [the prior art] met the limitations of the
claims.”).
IPR2020-00052
Patent 8,139,651 B2

30
variation in visual elements across adjacent pixels.” Pet. Reply 6 (citing Ex.
2049, 143:13–17), 10.
For the same reasons, we are also not persuaded by Petitioner’s
arguments that Vehviläinen teaches calculating level of detail according to
the typical usage of the term (id. at 6) or that variance, min-max, and SAD
calculations are all valid ways to assess level of detail that were used in the
art for the same purpose (id. at 9). On the contrary, the hypotheticals
presented by the parties highlight that, for a given set of pixel values,
Vehviläinen’s variance calculation can result in a very different calculation
than that resulting from the SAD embodiment disclosed in the ’651 patent,
thereby leading to a different filter choice for the same pixel set. Id. at 8–9.
Based on the evidence in the entire trial record, we find that Petitioner
has not established by a preponderance of the evidence that Vehviläinen
teaches or suggests the determining limitation of claim 1.
2. Conclusion on Anticipation and Obviousness of Claims 1,
17, 18, and 19 based on Vehviläinen
Based on the foregoing, we determine Petitioner has not shown by a
preponderance of the evidence that independent claim 1 or its dependent
claims 17 and 18 are anticipated by Vehviläinen.
“Once all relevant facts are found, the ultimate legal determination [of
obviousness] involves the weighing of the fact findings to conclude whether
the claimed combination would have been obvious to an ordinary artisan.”
Arctic Cat Inc. v. Bombardier Recreational Prods. Inc., 876 F.3d 1350, 1361
(Fed. Cir. 2017). On balance, considering the full record before us, we
determine that Petitioner has not established, by a preponderance of the
evidence, that Vehviläinen would have rendered the subject matter of
IPR2020-00052
Patent 8,139,651 B2

31
independent claim 1 or its dependent claims 17–19 obvious to one of
ordinary skill in the art at the time of the invention.10 Mylan Pharms., 914
F.3d at 1376 (“Dependent claims, with added limitations, are generally not
obvious when their parent claims are not.”).
F. Obviousness over Vehviläinen and Kadono
Petitioner contends that claims 1, 2, 4, and 17–19 are unpatentable
under 35 U.S.C. § 103 as obvious over Vehviläinen and Kadono. Pet. 52–
86. For the reasons that follow, we are not persuaded that Petitioner has
established by a preponderance of the evidence that claims 1, 2, 4, and 17–
19 are unpatentable under § 103(a) in view of Vehviläinen and Kadono.
1. Independent Claim 1
Patent Owner argues that the combination of Vehviläinen and Kadono
fails to teach or suggest the “determining the level of detail” limitation of
independent claim 1. PO Resp. 14–22. We, therefore, begin our discussion
with the parties’ arguments on this limitation.

10 Given our determination that Petitioner has failed to establish, by a
preponderance of the evidence, that Vehviläinen teaches or suggests all of
the limitations of the only independent claim of the ’651 patent, we need not
assess Patent Owner’s objective evidence of non-obviousness. See, e.g.,
Hamilton Beach Brands, Inc. v. f’real Foods, LLC, 908 F.3d 1328, 1343
(Fed. Cir. 2018) (holding, in affirming a Board decision determining that
petitioner had not shown unpatentability, that “objective indicia of
nonobviousness” “need not [be] addressed” because the court “affirmed the
Board’s findings regarding the failure of the prior art to teach or suggest all
[claim] limitations”); see also Mylan Pharm. Inc. v. Research Corp. Techs.,
Inc., 914 F.3d 1366, 1376 (Fed. Cir. 2019).
IPR2020-00052
Patent 8,139,651 B2

32
a) “determining the level of detail of the reconstructed
video frame across a region in which the block boundary
is located, wherein the region includes pixels from
multiple rows and multiple columns of the reconstructed
video frame that encompass pixels immediately adjacent
to at least two sides of the block boundary and includes
at least one pixel that is not immediately adjacent to the
block boundary”
(1) The Petition
Petitioner contends that both Vehviläinen and Kadono teach this
limitation and, when combined, the level of detail would be calculated using
a sum of absolute differences between pixels to determine whether to apply
deblocking to an area on both sides of a block boundary. Pet. at 64 (citing
Ex. 1005 ¶ 203; Ex. 1003 ¶ 156). Petitioner argues that the combined
teachings of Vehviläinen and Kadono render obvious the step of determining
the level of detail of the reconstructed video frame across a region, i.e., the
pixels marked by dots in Vehviläinen Figure 5A. Id.
Petitioner further argues the combination of Vehviläinen and Kadono
renders this claim limitation obvious based on the application of Kadono’s Ω
threshold to Vehviläinen step 80 (of Figure 6) by including an additional
comparison to Ω for selecting between the “smooth” filter and a stronger
filter. Id. at 65 (citing 1003 ¶ 157). For that second comparison, Petitioner
contends that an ordinarily skilled artisan could have also used the area
marked by pixel groups r1, r2, r3, and r4 in Figure 19 of Kadono
(reproduced above) instead of the region shown in Figure 5A of
Vehviläinen. Id. (citing Ex. 1005, Fig. 19; Ex. 1003 ¶ 157). According to
Petitioner, Kadono’s 4 x 4 pixel region comprising four adjacent column
regions r3, r1, r2, and r4 also teaches the claimed region. Id. at 66–67.
IPR2020-00052
Patent 8,139,651 B2

33
Petitioner argues that a person of ordinary skill in the art would have
been motivated to combine the teachings of Vehviläinen with Kadono by
starting with Vehviläinen’s teachings regarding adaptive deblocking and
applying Kadono’s calculation of pixel variation using the sum of absolute
differences or SAD. Pet. 52 (citing Ex. 1003 ¶ 141). Petitioner argues that
Vehviläinen and Kadono teach similar deblocking techniques, both
determine the level of detail by calculating pixel variance for two-
dimensional areas around block boundaries to select appropriate filters, and
both are designed to perform the same function (deblocking) for commonly
used video standards. Id. at 52–53 (citing Ex. 1004, 2:19–25, 11:29–55,
9:51–67, 4:26–27, Fig. 5A; Ex. 1005 ¶¶ 8, 13, 125, 142, 205, 206, Fig. 19;
Ex. 1003 ¶¶ 141, 142). Petitioner argues that given the similarities between
the two references, a person of ordinary skill in the art would have
recognized the compatibility between them and combined their teachings.
Id. at 53–54. Petitioner asserts that “[i]t was common in the art of video
processing to combine various filtering techniques, and a POSTIA would
have regarded the combination of teachings from Vehviläinen and Kadono
as typical in the field.” Id. at 53 (citing Ex. 1003 ¶ 142). Petitioner also
asserts that “[i]t was common in the art of video encoding to use different
settings and tweaks to optimize visual quality and performance for particular
applications, computer platforms, and types of content.” Id. at 54 (citing Ex.
1003 ¶ 143; Ex. 1004, 12:13–16; Ex. 1005 ¶¶ 282–283).
Specifically as to the use of Kadono’s SAD calculation, Petitioner
argues that a person of ordinary skill in the art would have found it obvious
to apply that calculation to Vehviläinen’s method by replacing Vehviläinen’s
variance calculations (equation (4)) with Kadono’s teachings regarding
IPR2020-00052
Patent 8,139,651 B2

34
calculating the sum of absolute values of differences between pairs of pixels
in corresponding columns across the block boundary to form a
representative difference, and then averaging the representative differences
between column groups together. Pet. 54 (citing Ex. 1004, 10:44–55; Ex.
1005 ¶¶ 205, 206, Fig. 19; Ex. 1003 ¶ 144). Petitioner contends that this
would have been a simple substitution of one known element for another,
with predictable results. Id. at 54–55 (citing Ex. 1003 ¶ 144).
Petitioner provides two reasons for the proposed SAD substitution.
Petitioner argues that an ordinarily skilled artisan would have been
motivated to make the proposed replacement because SAD was a widely
used technique for calculating the level of detail and was widely recognized
as an accurate and effective method for analyzing pixel variation. Id. at 55
(citing Ex. 1009 ¶¶ 55–65; Ex. 1003 ¶ 145). Petitioner further argues that by
2004, single instruction multiple data (“SIMD”) processors were common in
the industry, and Kadono’s teaching of using such processors to accelerate
SAD calculations would have motivated an ordinarily skilled artisan to
improve Vehviläinen’s method by applying Kadono’s teachings, particularly
for use on SIMD computer platforms. Id. (citing Ex. 1005 ¶¶ 204, 205; Ex.
1003 ¶ 145).
(2) Patent Owner’s Response
Patent Owner responds that Petitioner’s theory directly contradicts
Vehviläinen’s express teachings and fails to provide any reason why a
person of ordinary skill in the art would have even attempted the proposed
modification. PO Resp. 14–15, 23 (citing Ex. 2050 ¶ 72). Patent Owner
argues that Petitioner offers no analysis or other evidence showing that
Kadono’s SAD calculation was superior to Vehviläinen’s variance method.
IPR2020-00052
Patent 8,139,651 B2

35
Id. at 24. On the contrary, Patent Owner contends, Dr. Bove admitted that
Vehviläinen’s unmodified variance calculation was already “accurate and
effective.” Id. (citing Ex. 2049, 144:2–8, 149:4–9; Ex. 2050 ¶¶ 74–75).
Patent Owner further contends that aside from allowing the use of SIMD
processors, Kadono does not teach any benefit to using its SAD calculation.
Id. at 25 (citing Ex. 2049, 163:8–23). Patent Owner argues that “in addition
to Vehviläinen’s aversion to focusing on the differences between adjacent
pixels, . . . Vehviläinen expressly seeks to avoid even the general type of
filtering (low pass filtering) that both the [’651 patent] and Kadono utilize.”
Id. at 26 (citing Ex. 1001, 10:45; Ex. 1005 ¶ 132; Ex. 1004, 2:21–25, 6:40–
45). According to Patent Owner, “[e]ach of Vehviläinen and Kodono
utilizes the calculation that best suits its own subsequent filtering process,
each of which is, according to each reference ostensibly complete,” and an
ordinarily skilled artisan would not be motivated to combine two references
that independently operate effectively to accomplish similar functions. Id. at
27–28 (citing Ex. 2050 ¶ 77; Kinetic Concepts, Inc. v. Smith & Nephew, Inc.,
688 F.3d 1342, 1369 (Fed. Cir. 2012)).
With regard to Petitioner’s claimed motivation based on the
implementation of Kadono’s SAD calculation on SIMD processors, Patent
Owner argues that the benefits of using SIMD to speed up calculation would
have worked just as well to speed up Vehviläinen’s existing variance-based
calculations. PO Resp. 28–29 (citing Ex. 2050 ¶¶ 82, 83; Ex. 2016, 64
(describing variance calculations in the very first SIMD supercomputer in
1975)). Patent Owner contends that Petitioner offers no basis to conclude
that any benefits from using SIMD would have been superior with SAD as
opposed to variance such that it would motivate an ordinarily skilled artisan
IPR2020-00052
Patent 8,139,651 B2

36
to modify Vehviläinen to use SAD instead of its own variance. Id. at 29.
Patent Owner asserts that SIMD and its utility were known and widely
available at the time of Vehviläinen, which nonetheless chose to utilize
variance calculations, not SAD. Id. at 30 (citing Ex. 2017, 2; Ex. 2018, 1;
Ex. 2019, 7) (discussing SIMD architecture in processors from Hewlett-
Packard, Sun Microsystems, and Intel). Moreover, Patent Owner asserts, the
use of SAD instructions in video compression and its implementation in
SIMD processors were well known in the art before Vehviläinen. Id. at 30–
31 (citing Ex. 2020, 57, 58). Patent Owner argues that Vehviläinen used
variance not because SAD techniques were unavailable or unknown to
Vehviläinen, but because Vehviläinen’s filtering analysis is designedly
based on analysis of pixels in comparison to an average of all pixels and not
to adjacent pixels. Id. (citing Ex. 2050 ¶ 85).
(3) Petitioner’s Reply
In its Reply, Petitioner argues that Patent Owner’s requirement that
the determining the level of detail step must be tailored to the filter selection
step is contradicted by the ’651 patent and Vehviläinen, which both suggest
flexibility in swapping out components. Pet. Reply 21. Petitioner asserts
that the ’651 patent mentions that where the level of detail calculation
indicates a smooth region, the MPEG-4 DC offset filter can be used, and that
other smoothing filters can be applied. Id. (citing Ex. 1001, 10:12–19).
Similarly, Petitioner asserts, Vehviläinen interchanges variance or min-max
for calculating level of detail. Id. at 22 (citing Ex. 1004, 10:32–55).
Petitioner relies on Dr. Bove’s testimony that SAD “would yield a
classification of the same sort” as Vehviläinen’s variance calculation, and
that “[b]oth of these techniques were well known in the art many years prior
IPR2020-00052
Patent 8,139,651 B2

37
. . . so the classification technique would be regarded as interchangeable
. . . in terms of building an adaptive deblocking system,” to argue that a
person or ordinary skill in the art would be motivated to apply Kadono’s
SAD calculation to Vehviläinen. Id. (quoting Ex. 2049, 183:24–184:23; Ex.
1003 ¶¶ 144–146).
Petitioner also argues, for the first time in its Reply, that a person of
ordinary skill in the art “would have been motivated to apply [Kadono’s
SAD calculation] when using computers with special SIMD instructions for
SAD—this was common given the popularity of the technique.” Id.
(emphasis added) (citing Ex. 1020, 5). According to Petitioner, “[t]here are
certain SIMD instruction sets that explicitly can calculate sum of absolute
difference in parallel on multiple pixels at on[c]e . . . those hardware
platforms might not be able to implement the comparisons needed to do
min-max as efficiently.” Id. (quoting Ex. 2049, 182:13–183:17).
Specifically, Petitioner points to the PSADBW instruction in Intel’s SIMD
extensions that “calculate[] the sum of the absolute values of the differences
between two packed unsigned byte integer vectors . . . in one step.” Id. at
22–23 (citing Ex. 1020, 5).
(4) Analysis
On consideration of the full record, we determine that Petitioner does
not articulate an adequate motivation to combine Vehviläinen and Kadono to
arrive at the claimed invention. As discussed above, we are persuaded by
Patent Owner’s argument that Vehviläinen’s variance calculation is tied
closely to its filter design, which seeks to maintain differences between
adjacent video pixels, and therefore does not look to differences between
adjacent pixels in determining the level of detail for filter selection.
IPR2020-00052
Patent 8,139,651 B2

38
See supra § III.E.1.a.4. Vehviläinen’s stated goal is to reduce blocking
artifacts without significantly degrading the picture resolution, which it
seeks to realize with a filter that shifts pixel values toward a reference line
and “not by low-pass filtering.” Ex. 1004, 2:18–24; see also id. at 6:40–45
(“The main idea of this invented filter is to reduce blocking artefacts without
significantly degrading the picture details.”), 1:61–67 (“One significant
disadvantage of filtering is that the sharpness of the video picture is
reduced.”), code (57) (“If the filtering according to the invention is
performed, it is focused to a certain number of bits . . . close to the boundary
[and] is done by adjusting the numerical values of each video pixel close to
the boundary towards a reference line . . . .”).
Vehviläinen repeatedly makes the point that its filtering method works
better at preserving picture resolution because it does not factor in
differences between adjacent pixels:
This reduces the disappearance of the resolution in the picture in
the filtering process according to the invention because the
adjacent video pixels as such are not filtered in comparison with
each other but are adjusted in comparison to the reference line.
Id. at 2:41–45; see also id. 8:63–9:6. Dr. Bajaj’s testimony that
“Vehviläinen deliberately sets out to avoid founding its filtering scheme
upon the differences in the values of adjacent pixels” is therefore persuasive.
Ex. 2050 ¶¶ 61, 62 (citing Ex. 1004, 9:51–67; 10:32–44). Vehviläinen’s
filter selection is tied directly to its variance calculation and “uses three
branches (smooth, moderate, and high activity) according to the activity
inside adjacent blocks and along the boundary between two adjacent
blocks.” Ex. 1004, 6:46–49, 10:1–31 (detailing Vehviläinen’s branch
selection process). Vehviläinen discloses that the manner in which the
IPR2020-00052
Patent 8,139,651 B2

39
reference line is defined and number of pixels that are moved toward the
reference line depend directly upon the filter selection, i.e., the variance
determination. Id. at 7:11–8:10; Figs. 3A–3F, 4. Thus, replacing
Vehviläinen’s variance calculation with Kadono’s SAD calculation to select
a filter based on the differences between adjacent pixels rather than variance
across the entire region would have been understood by an ordinarily skilled
artisan as changing the basic principle of Vehviläinen’s operation,
consequently dissuading the combination with Kadono. See In re Ratti, 270
F.2d 810, 813 (CCPA 1959) (holding that a proposed combination of
references is not proper where it would change the basic principles of
operation of one of the references); Plas-Pak Indus., Inc. v. Sulzer Mixpac
AG, 600 F. App’x. 755, 759 (Fed. Cir. 2015) (“a change in a reference’s
‘principle of operation’ is unlikely to motivate a person of ordinary skill to
pursue a combination with that reference”). A person of ordinary skill in the
art would have no motivation to make Petitioner’s proposed modification to
Vehviläinen because that person would have also understood the two
calculation methods to be considerably different as to have a significant
impact on Vehviläinen’s filter operation, defeating its stated goal of reducing
blocking artefacts without significantly degrading the picture details. See
e.g., Ex. 1004, 2:18–24.
Petitioner’s reliance on the ’651 patent as “suggest[ing] flexibility in
swapping out components” is also misplaced. Pet. Reply 21–22 (citing
Ex. 1001, 10:12–19). First, that general statement from the ’651 patent
Specification that “other smoothing filters can be applied” in other
embodiments does not suggest that any smoothing filter can be plugged into
the disclosed embodiment. Second, relying on the challenged patent in
IPR2020-00052
Patent 8,139,651 B2

40
determining whether a person of ordinary skill would have been led to make
the proposed modification requires use of hindsight. See Cardiac
Pacemakers, Inc. v. St. Jude Medical, Inc., 381 F.3d 1371, 1376 (Fed. Cir.
2004) (“the suggestion to combine references must not be derived by
hindsight from knowledge of the invention itself.”); Otsuka Pharma. Co., v.
Sandoz, Inc., 678 F.3d 1280, 1296 (Fed. Cir. 2012) (“The inventor’s own
path itself never leads to a conclusion of obviousness; that is hindsight.”).
Petitioner also points us to the fact that Vehviläinen has at least two
different methods of variance calculation as suggesting flexibility in
swapping out components. Id. at 22 (citing Ex. 1004, 10:32–55). Those two
different methods, however, arrive at the same result, i.e., variance.
Ex. 1004, 10:32–55 (“variance can be estimated by a more simple method
called min-max approximation”); Ex. 2049, 149:4–9; Ex. 2050 ¶ 67. That
Vehviläinen provides a simpler method to calculate the same property, i.e.,
variance, does not support Petitioner’s contention that it provides the
flexibility to change its filter selection decision based on an entirely different
determination. See Adidas AG v. Nike, Inc., 963 F.3d 1355, 1359 (Fed. Cir.
2020) (“Fundamental differences between the references are central to th[e]
motivation to combine inquiry.”).
Dr. Bove’s testimony that “[t]he substitution of one for the other
would not have changed the principle of operation for either reference
because Vehviläinen and Kadono use similar mechanisms for a similar
purpose,” is unpersuasive. See Ex. 1003 ¶ 144. As Dr. Bove himself
explains, “one of ordinary skill in the art in making that choice would want
to have assurance that either determining step being considered would
generally classify high-activity blocks in the same way, despite doing a
IPR2020-00052
Patent 8,139,651 B2

41
different calculation.” Ex. 2049, 169:6–1 (emphasis added); see also id. at
170:16–19 (“Now, in so doing, one of ordinary skill in the art would read the
teachings and would say, How is this thing described.”). Yet, Petitioner and
Dr. Bove fail to reconcile the fundamental differences in the filtering
techniques of Vehviläinen and Kadono. Instead, Dr. Bove disregards such
concerns as “not a significant barrier” because “that sort of thing was well
known in the art, much prior to the period of time we’re talking about here.”
Id. at 169:12–15. We are not persuaded that an ordinarily skilled artisan,
mindful of how Vehviläinen’s filter mechanism works, would be motivated
to replace its variance calculation with Kadono’s SAD calculation merely
because the two were well known in the art, especially given that such a
person would understand Vehviläinen variance calculation “to be an
accurate and effective way to determine the activity inside a block.” Id. at
144:2–8. In view of the inconsistencies between Dr. Bove’s testimony and
the teachings of Vehviläinen, we do not give substantial weight to his
opinion.
For the same reasons, we are not persuaded that an ordinarily skilled
artisan would have been motivated to apply Kadono’s teachings to improve
Vehviläinen’s techniques based on Kadono’s use of SIMD processors. Pet.
55. Because we determine that the use of Kadono’s SAD calculation in
Vehviläinen’s system would change its basic operation and prevent
Vehviläinen’s filter from operating in its desired manner, Kadono’s use of
SIMD processors for SAD calculation would not have motivated a person of
ordinary skill in the art to switch Vehviläinen’s variance calculation.11 Here

11 We consider Petitioner’s Reply argument that a POSITA would have been
motivated to the combination by special SIMD computer instructions for
IPR2020-00052
Patent 8,139,651 B2

42
too, we credit Dr. Bajaj’s testimony that a person of ordinary skill in the art
would not have altered Vehviläinen’s variance calculation and filtering
mechanism in view of Kadono. See Ex. 2050 ¶¶ 81–85 (“These techniques
were widely known, yet Vehviläinen nevertheless utilized variance-based
calculations, not because SAD techniques were unavailable or unknown to
Vehviläinen, but because Vehviläinen’s filtering analysis is designedly
based on analysis of pixels in comparison to an average of all pixels and not
to adjacent pixels.”).
Based on the evidence in the entire trial record, we find that Petitioner
has not articulated an adequate motivation to combine Vehviläinen and
Kadono to arrive at the claimed invention.
2. Conclusion on Obviousness of Claims 1, 2, 4, 17–19 based
on Vehviläinen and Kadono
“Once all relevant facts are found, the ultimate legal determination [of
obviousness] involves the weighing of the fact findings to conclude whether
the claimed combination would have been obvious to an ordinary artisan.”
Arctic Cat, 876 F.3d at 1361. On balance, considering the full record before
us, we determine that Petitioner has not established, by a preponderance of
the evidence, that the combination of Vehviläinen and Kadono would have
rendered the subject matter of independent claim 1 or its dependent claims 2,

SAD (Pet. Reply 22–23) to be forfeited. See 37 C.F.R. § 42.23(b); see also
Intelligent Bio-Sys., 821 F.3d at 1369 (emphasizing “the requirement that the
initial petition identify ‘with particularity’ the ‘evidence that supports the
grounds for the challenge to each claim.’”). Even if not forfeited, we are not
persuaded that such instructions support a motivation to combine given our
determination that use of Kadono’s SAD calculation would change
Vehviläinen’s basic principle of operation.
IPR2020-00052
Patent 8,139,651 B2

43
4, and 17–19 obvious to one of ordinary skill in the art at the time of the
invention. 12
IV. PATENT OWNER’S MOTION TO EXCLUDE
Patent Owner filed a Motion to Exclude seeking to exclude Exhibits
1019–1023, and 1026. Paper 72 (PO Mot.). Petitioner filed an opposition
(Paper 74 (“Pet Opp.”)) and Patent Owner filed a reply (Paper 78 (“PO
Reply”)). Patent Owner, as the moving party, has the burden of proof to
establish that it is entitled to the requested relief. 37 C.F.R. §§ 42.20(c),
42.62(a).
A. Exhibit 1019
Patent Owner’s objections to Exhibit 1019 are based on Federal Rules
of Evidence (“FRE”) 901 (authentication) as well as FRE 801 and 802
(hearsay). Mot. 1, 2. Patent Owner argues that Petitioner has failed to offer
evidence that Exhibit 1019 accurately reproduces the Google searches that
Petitioner purports were performed. Id. at 1 (citing F.R.E. 901(a)). Patent
Owner contends that Exhibit 1019 is not a self-authenticating document. Id.
Patent Owner further argues that the exhibit contains inadmissible hearsay.
Id. at 2.
Petitioner responds that Exhibit 1019 is rebuttal evidence to Patent
Owner’s Exhibit 2053, and that if Patent Owner’s exhibit is admitted into
evidence, Exhibit 1019 should also be admitted. Pet. Opp. 1.

12 Given our determination that Petitioner has failed to establish, by a
preponderance of the evidence, an adequate motivation to combine
Vehviläinen and Kadono, we need not assess Patent Owner’s objective
evidence of non-obviousness.
IPR2020-00052
Patent 8,139,651 B2

44
Patent Owner responds that Exhibit 2053 is evidence of Google
searches that Dr. Bajaj performed, and that Petitioner has not moved to
exclude it. PO Reply at 1 (citing Ex. 2050 ¶ 29).
Exhibit 1019 appears to include printouts of the results of Google
searches conducted for the terms “the art of deblocking” and “the field of
deblocking,” without quotation marks. Petitioner, however, offers no
evidence to show that the exhibit is in fact a printout of such search results.
See FRE 901(a) (“To satisfy the requirement of authenticating or identifying
an item of evidence, the proponent must produce evidence sufficient to
support a finding that the item is what the proponent claims it is.”). We
therefore grant Patent Owner’s motion to exclude Exhibit 1019 for lack of
authentication.
Petitioner further asks that we take judicial notice of these Google
searches and results given that “[t]he facts in Exhibit 1019 are not subject to
reasonable dispute because DivX concedes to the propriety of Google
searches by submitting its Exhibit 2053.” Opp. 2. Given our decision above
regarding the level of ordinary skill in the art (supra § III.B) and that we do
not find recently conducted Google searches to be probative of the level of
ordinary skill in the art at the time of the invention, we are not persuaded to
grant Petitioner’s request for judicial notice.
B. Exhibit 1020
Patent Owner’s objections to Exhibit 1020 are based on FRE 801 and
802 (hearsay), as well as 403 (relevance). Mot. 3–5. Patent Owner also
argues that Exhibit 1020 should be excluded as untimely because it is
submitted in “support of Petitioner’s prima facie case of obviousness to
show alleged motivation to combine the reference.” Id. at 4.
IPR2020-00052
Patent 8,139,651 B2

45
We need not assess the merits of Patent Owner’s objections because,
even considering Petitioner’s evidence, we have determined that Petitioner
has not demonstrated by a preponderance of the evidence that the challenged
claims are unpatentable. Accordingly, Patent Owner’s objections are
dismissed as moot.
C. Exhibits 1021–1023
Patent Owner’s objections to Exhibits 1021–1023 are based on FRE
801 and 802 (hearsay), 403 (relevance), and 901 (authentication). Mot. 5–6.
Petitioner relies on these exhibits in support of its rebuttal to Mr. Hanson’s
declaration submitted by Patent Owner to establish secondary
considerations. Id. at 1.
Under the particular circumstances in this case, we need not assess the
merits of Patent Owner’s objections. As discussed above, we do not assess
Patent Owner’s objective evidence of non-obviousness and therefore do not
consider Petitioner’s rebuttal evidence. Accordingly, Patent Owner’s
objections to these exhibits are dismissed as moot.
D. Exhibits 1026
Patent Owner’s objections to Exhibit 1026 are based on FRE 801 and
802 (hearsay) and 403 (relevance). Mot. 7–8. Patent Owner contends that
Exhibit 1026 is a declaration made in a different proceeding and is therefore
out-of-court testimony in this proceeding. Id. at 7. Patent Owner further
contends that Petitioner uses the exhibit to prove that no testing would be
required for a POSITA to combine various aspects of codecs, thereby
relying on it for the truth of the matter asserted therein. Id. (citing Pet.
Reply 19–20). Patent Owner argues that there are dispositive differences
between combination of references here and the combination of references
IPR2020-00052
Patent 8,139,651 B2

46
in the other proceeding that Dr. Bajaj opines on in Exhibit 1026, thereby
undermining its probative value, especially when Patent Owner is not
allowed to submit rebuttal evidence to it. Id. at 7–8.
Petitioner responds that Exhibit 1026 is not used to show that
Dr. Bajaj’s prior testimony is correct, it is merely relied upon to show that
Dr. Bajaj’s testimony in the current proceeding is contradicted by his sworn
testimony in other proceedings before the Board. Opp. 9–10. According to
Petitioner, Dr. Bajaj’s testimony here that video encoding is an experimental
art and testing is necessary is contradicted by his opinions found in Exhibit
1026. Id. at 10 (citing Ex. 1017, 128:16–25, 126:2–6; Ex. 1026, ¶¶ 79, 67–
82). Petitioner asserts that Dr. Bajaj himself admitted that the other
proceeding challenged a similar patent to the one at issue here. Id. at 10–11
(citing Ex. 1017, 172:18–173:7).
Although we agree with Patent Owner that there are differences
between the facts presented in the two cases, we are not persuaded that the
probative value of the exhibit is outweighed by the risk of confusion or
unfair prejudice to Patent Owner. Moreover, because Petitioner offers the
exhibit to show a supposed contradiction between Dr. Bajaj’s testimony, we
do not consider the statements in Exhibit 1026 for the truth of the matter
asserted; instead, we consider it proper impeachment evidence. We
therefore deny Patent Owner’s motion to exclude.
A. Conclusion
Patent Owner has satisfied its burden to show that Exhibit 1019
should be excluded but has not satisfied that burden as to Exhibit 1026.
Patent Owner’s objections to Exhibits 1020–1023 are dismissed as moot.
IPR2020-00052
Patent 8,139,651 B2

47
Accordingly, Patent Owner’s Motion to Exclude is granted in part, denied in
part, and dismissed in part.
V. PETITIONER’S MOTION TO EXCLUDE
Petitioner filed a Motion to Exclude seeking to exclude Patent
Owner’s Exhibits 2051, 2054–2069, and 2073–2080. Paper 73 (“Pet Mot.”).
Patent Owner filed an opposition (Paper 76 (“PO Opp.”)) and Petitioner
filed a reply (Paper 77). Petitioner has the burden of proof to establish that it
is entitled to the requested relief. 37 C.F.R. §§ 42.20(c), 42.62(a).
A. Exhibits 2051, 2054–2069
Petitioner’s objections to Exhibits 2054–2068 are based on FRE
901(a) (authentication), 801(c) and 802 (hearsay), 701 (improper opinions),
and 402 (relevance). Pet. Mot. 2–8. Petitioner objects to portions of Exhibit
2051, which is the declaration of Seth Hanson, based on FRE 701 (improper
opinions), 602 (lack of personal knowledge), and because those portions rely
on 2054–2068. Id. at 12–14. Petitioner objects to Exhibit 2069, which Mr.
Hanson relies on in his declaration, based on FRE 901(a) (authentication),
801(c) and 802 (hearsay), and 701 (improper opinions). Id. at 14–15. Patent
Owner relies on these exhibits to establish secondary considerations.
Under the particular circumstances in this case, we need not assess the
merits of Petitioner’s objections. As discussed above, we do not consider
Patent Owner’s evidence of secondary considerations. Accordingly,
Petitioner’s objections to these exhibits are dismissed as moot.
B. Exhibits 2073–2079
Petitioner’s objections to Exhibits 2073–2079 are based on FRE
901(a) (authentication), 801(c) and 802 (hearsay), 701 (improper opinions),
and 402 and 403 (relevance). Pet. Mot. 8–12. Petitioner argues that these
IPR2020-00052
Patent 8,139,651 B2

48
webpages are unreliable and Patent Owner has not produced evidence
sufficient to support a finding that these exhibits are in fact published
articles and not fake or manipulated webpages. Id. at 8–9. Petitioner further
argues that these exhibits contain out-of-court statements by third parties and
that Patent Owner relies on those statements for the truth of the matters
asserted therein. Id. at 9. Petitioner also argues that Patent Owner uses the
exhibits for baseless character attacks, making them precisely the type of
irrelevant evidence that courts have excluded because of its tendency to
suggest a decision on an improper basis. Id. at 10. Petitioner contends that
there is no question that Dr. Bove is an expert in the field, and Exhibits
2073–2079 are not related to Dr. Bove’s expertise or the accuracy of his
expert opinions. Id. (citing Ex. 1017, 102:15–24; PO Resp. 4). Petitioner
further contends that none of these exhibits relate to Dr. Bove’s credibility or
truthfulness because they are entirely consistent with what Dr. Bove has
readily explained. Id. at 11 (citing Ex. 1003 ¶ 4; Ex. 2078, 1).
Patent Owner responds that these exhibits were authenticated with a
declaration providing evidence of their source, and points to various
characteristics of the exhibits that it contends make them self-authenticating.
Id. at 9–10. Patent Owner further argues that these exhibits are not hearsay
because they are not offered for the truth of the matter asserted but to
impeach Dr. Bove’s credibility and truthfulness. Id. at 10 (citing FRE
608(b)(1)).
We are persuaded that Exhibits 2073–2077 contain inadmissible
hearsay and are irrelevant to any of the issues in this proceeding. Patent
Owner relies on these exhibits in support of assertions unrelated to
IPR2020-00052
Patent 8,139,651 B2

49
Dr. Bove’s testimony, not to impeach Dr. Bove’s testimony. See PO Resp.
10–11. We therefore grant Petitioner’s motion as to these exhibits.
Patent Owner does rely on Exhibits 2078 and 2079 for impeachment
purposes arguing that those exhibits contradict certain testimony provided by
Dr. Bove. Id. at 11 (citing Ex. 2049, 173:11–20; 174:13–19). We find these
exhibits authentic and relevant for impeachment of the cited testimony, and
consider them for the limited purpose of judging Dr. Bove’s credibility. We
deny Petitioner’s motion as to the use of these exhibits for that limited
purpose.
C. Exhibit 2080
Petitioner’s objections to Exhibit 2080 are based on FRE 901(a)
(authentication), 801(c) and 802 (hearsay), and 701 (improper opinions).
Pet. Mot. 14–15. Petitioner argues that there is no evidence that this exhibit
was published during the 2004–2005 timeframe or that it is an accurate,
unmodified copy of the purported university thesis. Id. at 14. Petitioner
further argues that the exhibit is offered as hearsay because Patent Owner
relies on it for the truth of purported test results and analysis therein to
support “unpredictability of various combinations in the field [of
deblocking].” Id. at 15 (citing Ex. 2050 ¶ 40). Petitioner further argues that
the author of the exhibit has not been established as an expert and it should
be excluded as containing impermissible opinion testimony. Id.
Patent Owner responds that Exhibit 2080 has indicia of reliability
regarding its status as an accurate copy of a published university thesis,
including the university’s logo. PO Opp. 14. Patent Owner further argues
Exhibit 2080 is not hearsay because it is not offered for the truth of the
discoveries or conclusions reached in the thesis but show an ordinarily
IPR2020-00052
Patent 8,139,651 B2

50
skilled artisan’s understanding about unpredictability of combinations of
deblocking schemes. Id. at 15 (citing Ex. 2050 ¶ 40). Patent Owner also
contends that FRE 701 does not apply to this exhibit because it is not
testimonial evidence.
We agree with Patent Owner that Exhibit 2080 is not testimonial
evidence. Exhibit 2080 is one of the references that Dr. Bajaj relies upon in
support of his opinions relating to the state of art and the unpredictably of
combining deblocking techniques. See Ex. 2050 ¶ 40 (citing also Ex. 1007).
Exhibit 2080 is self-authenticating as it bears indicia of reliability regarding
its status as a thesis published by Eindhoven University. See Ex. 2080
(showing university’s logo, publication date, author’s name, supervisor’s
name, and a link to a copy the thesis available on the university’s website).
Moreover, Petitioner cross-examined Dr. Bajaj about this exhibit. See Ex.
1017, 115:2–119:8. FRE 703 allows experts to base their opinions on
inadmissible facts or data under certain circumstances and allows for such
evidence to be considered by the factfinder if its probative value
substantially outweighs their prejudicial effect. We are persuaded to allow
Exhibit 2080 for that limited purpose. Petitioner’s motion is denied as to
Exhibit 2080.
D. Conclusion
Petitioner has satisfied its burden to show that Exhibits 2073–2077
should be excluded, but has not satisfied that burden as to Exhibits 2078–
2080. Petitioner’s objections to Exhibits 2051 and 2054–2069 are dismissed
as moot. Accordingly, Petitioner’s Motion to Exclude is granted in part,
denied in part, and dismissed in part.
IPR2020-00052
Patent 8,139,651 B2

51
VI. PATENT OWNER’S CONSTITUTIONAL CHALLENGE
Patent Owner argues that “[t]he Board is unconstitutionally appointed
to decide IPRs,” and “[e]liminating Board tenure protections cannot cure
that problem.” PO Resp. 63 (citing Arthrex, Inc. v. Smith & Nephew, Inc.,
941 F.3d 1320 (Fed. Cir. 2019), cert. granted sub nom. United States v.
Arthrex, Inc., 2020 WL 6037206 (Oct. 13, 2020)). According to Patent
Owner, “[e]ven if a cure were possible without Congressional confirmation
or higher agency review, it cannot be done by ‘retroactively’ ‘severing’
long-guarded protections in a long-standing different statute.” Id. Patent
Owner further argues that “[b]ecause Patent Owner is not before the Board
by choice, and no properly-appointed Board panel is available, the
appropriate remedy is not severance, but dismissal.” Id.
We decline to consider Patent Owner’s constitutional challenge as the
issue has been addressed by the Federal Circuit in Arthrex, 941 F.3d at 1328.
VII. CONCLUSION
For the foregoing reasons, we conclude that Petitioner has not
established by a preponderance of the evidence that claims 1, 2, 4, 17–19 of
the ’651 patent are unpatentable.
IPR2020-00052
Patent 8,139,651 B2

52
In summary:
Claim(s) 35 U.S.C. § References/Basis
Claims
Shown
Unpatentable
Claims Not
Shown
Unpatentable
1, 17, 18 102(b) Vehviläinen 1, 17, 18
1, 17–19 103(a) Vehviläinen 1, 17–19
1, 2, 4,
17–19 103(a)
Vehviläinen,
Kadono 1, 2, 4, 17–19
Overall
Outcome 1, 2, 4, 17–19

VIII. ORDER
It is, therefore,
ORDERED that claims 1, 2, 4, and 17–19 of the ’651 patent have not
been shown to be unpatentable;
FURTHER ORDERED that Petitioner’s Motion to Exclude is granted
with regard to Exhibits 2073–2077, denied with regard to Exhibits 2078–
2080, and dismissed as moot with regard to Exhibits 2051 and 2054–2069;
FURTHER ORDERED that Patent Owner’s Motion to Exclude is
granted with regard to Exhibit 1019, denied with regard to Exhibit 1026, and
dismissed as moot with regard to Exhibits 1020–1023;
FURTHER ORDERED that, because this is a Final Written Decision,
parties to the proceeding seeking judicial review of the decision must
comply with the notice and service requirements of 37 C.F.R. § 90.2.

IPR2020-00052
Patent 8,139,651 B2

53
PETITIONER:

Harper Batts
Chris Ponder
Jeffrey Liang
SHEPPARD, MULLIN, RICHTER & HAMPTON, LLP
hbatts@sheppardmullin.com
cponder@sheppardmullin.com
jliang@sheppardmullin.com

PATENT OWNER:

Kenneth Weatherwax
Nathan Lowenstein
Bridget Smith
Parham Hendifar
Jason Linger
LOWENSTEIN & WEATHERWAX LLP
weatherwax@lowensteinweatherwax.com
lowenstein@lowensteinweatherwax.com
smith@lowensteinweatherwax.com
hendifar@lowensteinweatherwax.com
linger@lowensteinweatherwax.com