Google Inc.v.Network-1 Technologies, Inc.

Patent Trial and Appeal Board

Jun 20, 2016

11977202 (P.T.A.B. Jun. 20, 2016)

Trials@uspto.gov Paper 30
571-272-7822 Entered: June 20, 2016

UNITED STATES PATENT AND TRADEMARK OFFICE
____________
BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________
GOOGLE INC.
Petitioner,

v.

NETWORK-1 TECHNOLOGIES, INC.,
Patent Owner.
_______________

Case IPR2015-00345
Patent 8,205,237 B2
____________

Before KEVIN F. TURNER, LYNNE E. PETTIGREW, and
JON B. TORNQUIST, Administrative Patent Judges.

TURNER, Administrative Patent Judge.

FINAL WRITTEN DECISION
Inter Partes Review
35 U.S.C. § 318(a) and 37 C.F.R. § 42.73

I. INTRODUCTION
Google Inc. (“Petitioner”) filed a Petition (Paper 1, “Pet.”) requesting
inter partes review of claims 1–27, 29, 30, 32–35, 37, 38, and 40 of U.S.
Patent No. 8,205,237 B2 (Ex. 1001, “the ’237 Patent”). Network-1
IPR2015-00345
Patent 8,205,237 B2

2

Technologies, Inc. (“Patent Owner”) filed a Preliminary Response to the
Petition. Paper 5 (“Prelim. Resp.”). We instituted the instant trial on June
23, 2015, with respect to claims 1–16, 21–27, 29, 30, 33–35, 37, and 38,
pursuant to 35 U.S.C. § 314. Paper 6 (“Dec.”).
Subsequently, Patent Owner filed a Response (Paper 17, “PO Resp.”),
and Petitioner filed a Reply (Paper 20, “Reply”). Further to authorization
provided in a conference call, Patent Owner filed a paper identifying
allegedly improper arguments in the Reply (Paper 24), and Petitioner filed a
response thereto (Paper 25). Oral hearing1 was held on March 9, 2016, and
a transcript of the hearing was entered into the record. Paper 29 (“Tr.”).
We have jurisdiction under 35 U.S.C. § 6(c). This Final Written
Decision is entered pursuant to 35 U.S.C. § 318(a) and 37 C.F.R. § 42.73.
For the reasons set forth below, we determine that Petitioner has shown, by a
preponderance of the evidence, that claims 9–16, 23, and 24 of the ’237
Patent are unpatentable, but has not shown that claims 1–8, 21, 22, 25–27,
29, 30, 33–35, 37, and 38 of the ’237 Patent are unpatentable.
A. Related District Court Proceedings
The parties inform us that the ’237 Patent is the subject of the
following lawsuit: Network-1 Technologies, Inc. v. Google Inc. and
YouTube, LLC, Case No. 1:14-cv-02396 (S.D.N.Y.). Pet. 1. YouTube, LLC
is a subsidiary of Petitioner, and is acknowledged as a real party-in-interest.
Id. In addition, three additional patents, U.S. Patent Nos. 8,640,179,
8,010,988, and 8,656,441, all issuing from applications related to the ’237

1 The hearings for this review and IPR2015-00343, IPR2015-00347, and
IPR2015-00348 were consolidated.
IPR2015-00345
Patent 8,205,237 B2

3

Patent, are subject to inter partes reviews, namely IPR2015-00343,
IPR2015-00347, and IPR2015-00348, respectively.
B. The ’237 Patent
The ’237 Patent relates to identifying a work, such as a digital audio
or video file, without the need to modify the work. Ex. 1001, 1:31–36,
4:25–31. This identification can be accomplished through the extraction of
features from the work, and comparison of those extracted features with
records of a database or library. Id. at Abstract. Thereafter, an action may
be determined based on the identification determined. Id. at 4:24–25.
Figure 1, reproduced below, illustrates the steps of the claimed computer-
implemented methods:

Fig. 1 of the ’237 Patent illustrating the claimed method
IPR2015-00345
Patent 8,205,237 B2

4

C. Illustrative Claims
Claim 1 is independent, along with claims 5, 9, 13, 25, and 33.
Claims 1, 9, and 25 are considered representative of the claims challenged,
and are reproduced below:
1. A computer-implemented method comprising:
a) receiving, by a computer system including at least one
computer, features that were extracted from a media work by a
client device;
b) determining, by the computer system, an identification of the
media work using the received features extracted from the
media work to perform a sub-linear time search of extracted
features of identified media works to identify a neighbor; and
c) transmitting, by the computer system, information about the
identified media work to the client device.

9. A computer-implemented method comprising:
a) receiving, by a computer system including at least one
computer, features what were extracted from media work by a
client device;
b) determining, by the computer system, an identification of the
media work using the received features extracted from the
media work to perform an approximate nearest neighbor
search of extracted features of identified media works; and
c) transmitting, by the computer system, information about the
identified media work to the client device.

25. A computer-implemented method comprising:
a) obtaining, by a computer system including at least one
computer, media work extracted features that were extracted
from a media work, the media work uploaded from a client
device;
IPR2015-00345
Patent 8,205,237 B2

5

b) determining, by the computer system, an identification of the
media work using the media work extracted features to perform
a nonexhaustive search of reference extracted features of
reference media works to identify a near neighbor; and
c) determining, by the computer system, an action based on the
determined identification of the media work.
Ex. 1001, 25:2–13, 25:62–26:5, 27:11–22 (emphases added).
D. Prior Art Relied Upon
Based on the instituted grounds, Petitioner relies upon the following
prior art references:
Reference or Declaration Exhibit No.
U.S. Patent No. 7,444,353 (“Chen”) Ex. 1008
U.S. Patent No. 5,874,686 (“Ghias”) Ex. 1010
U.S. Patent No. 6,188,010 ("Iwamura") Ex. 1012
E. Grounds of Unpatentability
We instituted the instant trial based on the following grounds of
unpatentability (Dec. 21–22):
Claims Challenged Basis Reference(s)
1, 3–5, 7–9, 11–13, 15, 16, 21–25, 29,
30, 33, 37, and 38
§ 102(e) Iwamura
1–3, 5–7, 9–11, 13–15, and 21–24 § 102(b) Ghias
26, 27, 34, and 35 § 103 Iwamura and Chen

II. ANALYSIS
A. Claim Construction
In an inter partes review, claim terms in an unexpired patent are given
their broadest reasonable construction in light of the specification of the
IPR2015-00345
Patent 8,205,237 B2

6

patent in which they appear. 37 C.F.R. § 42.100(b); Cuozzo Speed Techs.,
LLC v. Lee, No. 15–446, slip op. at 13 (U.S. June 20, 2016).
Under the broadest reasonable construction standard, claim terms are
given their ordinary and customary meaning, as would be understood by one
of ordinary skill in the art in the context of the entire disclosure. In re
Translogic Tech., Inc., 504 F.3d 1249, 1257 (Fed. Cir. 2007). Any special
definition for a claim term must be set forth with reasonable clarity,
deliberateness, and precision. In re Paulsen, 30 F.3d 1475, 1480 (Fed. Cir.
1994).
In our Institution Decision, we construed the following claim terms
specifically:
Claim Term(s) Construction
“non-exhaustive
search”
“a search that locates a match without a comparison
of all possible matches”
“sub-linear search” “a search whose execution time scales with a less
than linear relationship to the size of the data set to
be searched”
“neighbor search,”
“near neighbor
search”
“identifying a close, but not necessarily exact or
closest, match”
“nearest neighbor
search”
“identifying a closest, but not necessarily exact
match”
“approximate
nearest neighbor
search”
“identifying a close match that is not necessarily the
closest match”
Dec. 5–9.
Petitioner and Patent Owner agree largely with the adopted
constructions. PO Resp. 2–11; Reply 2–6. Below, we discuss the points
raised by the parties with respect to portions of those constructions and how
those portions should be interpreted.
IPR2015-00345
Patent 8,205,237 B2

7

i) “sub-linear search”
With respect to the construction of “sub-linear search,” or “sublinear
search,” Patent Owner argues that there are two possible interpretations of
that construction, with the “size of the data set” being the records in the data
set being searched, or the length of an individual record in the database. PO
Resp. 2. Patent Owner argues that the evidence of the Specification of the
’237 Patent and the testimony of the declarants point to the size of the data
set being the number of records being searched. Id. at 3–6. Petitioner
responds that “size of the data set,” according to its plain meaning, would be
“the amount of disk space a data set occupies.” Reply 2. Petitioner argues
that the Specification of the ’237 Patent supports this interpretation. Id. at 2-
4 (citing Ex. 1001, 21:14–26; Ex. 1020, 104:13–105:8). We agree with
Patent Owner with respect to this claim construction.
Although the sections of the Specification of the ’237 Patent cited by
Petitioner relate to disk size, that size is a direct result of the number of
records in the data set. Claim 1, for example, recites “using the received
features extracted from the media work to perform a sub-linear time search
of extracted features of identified media works.” For that claim, under our
adopted claim construction, the size of the data set would be the number of
extracted features of the identified media works. It would not be logical to
assume that the size of the data set would be the length of an individual
record or the amount of disk space the data set occupies when the number of
extracted features would be readily available. Although disk space could be
used as a measure, it would be variable with respect to the format of the disk
space, for example. Therefore, we are not persuaded that disk space should
be used as a determinant of data set size in the construction of “sub-linear
IPR2015-00345
Patent 8,205,237 B2

8

search.” As such, the data set would be the number of extracted features, in
claim 1, irrespective of the exact number of commercials or frames.
ii) “non-exhaustive search”
With respect to the claim term “non-exhaustive search,” Patent Owner
emphasizes that a non-exhaustive search makes a comparison without all
possible matches but does not concern itself with whether all data within all
possible matches have been compared. PO Resp. 6–7. Petitioner does not
dispute the construction, and we remain persuaded that a “non-exhaustive”
search need not consider all data within all possible matches, but rather
should be exhaustive or non-exhaustive with respect to the number of
records in a database. See Dec. 5–7.
iii) “neighbor / near neighbor search”
Patent Owner also contends that if a search necessarily identifies an
exact or the closest match, it is not a neighbor or near neighbor search
because it is not a search that identifies a close, but not necessarily exact or
closest, match. PO Resp. 7–8. We disagree with Patent Owner that a
neighbor search could not identify an exact or the closest match and still be a
neighbor search. Patent Owner, in its Response, appears to take an
“exclusive” view of the construction, arguing that if an exact or closest
match happens to be obtained in a neighbor or near neighbor search, that
search could not be a neighbor or near neighbor search. As discussed at Oral
Hearing, Patent Owner agrees that if a neighbor search does produce an
exact match, that effect does not necessarily negate the type of search
performed. Tr. 57–60 (“MR. DOVEL: If you're doing a neighbor search,
one that it's looking and it happens to pull up an exact match, that would still
be a neighbor search”).
IPR2015-00345
Patent 8,205,237 B2

9

iv) “approximate nearest neighbor search”
With respect to the claim limitation “approximate nearest neighbor
search,” Patent Owner argues that our construction is incomplete because the
Specification of the ’237 Patent requires such a search to be sub-linear and a
search that does not always find the closest match. PO Resp. 8–11. Patent
Owner cites a section of the Specification reciting that “[o]ne example of a
sub-linear time search is an approximate nearest neighbor search . . . [which]
does not always find the closest point to the query,” such that if the closest
match is guaranteed, the search is not an approximate nearest neighbor
search. Id. at 8–9 (citing Ex. 1001, 9:12–19). Patent Owner also alleges that
we applied “faulty logic” in excluding any non-linear aspect from the claim
construction of “approximate nearest neighbor search.” Id. at 10.
Petitioner responds that the statements in the Specification do not rise
to the level of lexicography or disavowal. Reply 4–5 (citing Facebook, Inc.
v. Pragmatus AV, LLC, 582 F. App'x 864, 868 (Fed. Cir. 2014); Renishaw
PLC v. Marposs Societa' per Azioni, 158 F.3d 1243, 1249 (Fed. Cir. 1998)).
We agree, given that the cited portion provides that an approximate nearest
neighbor search may be an example of a sublinear search, but not that all
approximate nearest neighbor searches must be sub-linear. Additionally,
Patent Owner has not provided persuasive evidence that all approximate
nearest neighbor searches must be sub-linear, other than the citation to the
Specification. As well, similar to our discussion of neighbor searches above,
we are not persuaded that if an approximate nearest neighbor search happens
to arrive at an exact match, that would negate the nature of the search. We
are persuaded that an approximate nearest neighbor search would depend on
the search methodology and not on the specific results obtained.
IPR2015-00345
Patent 8,205,237 B2

10

Additionally, as Petitioner points out, the instant claims confirm that
sublinearity is not necessarily required for an approximate nearest neighbor
search. Reply 5–6. Claim 33 recites, in part, a “sublinear approximate
nearest neighbor search,” whereas claims 9 and 12 simply recite an
“approximate nearest neighbor search,” without the sublinear modifier.
Interpreting all recitations of approximate nearest neighbor searches as
intrinsically sublinear would render the limitation in claim 33 superfluous.
Id. (citing Phillips v. AWH Corp., 415 F.3d 1303, 1314 (Fed. Cir. 2005);
Digital-Vending Servs. Int’l, LLC v. Univ. of Phoenix, Inc., 672 F.3d 1270,
1274–75 (Fed. Cir. 2012)). For this additional reason, we remain persuaded
that an approximate nearest neighbor search corresponds to “identifying a
close match that is not necessarily the closest match,” without a requirement
of being sublinear.
Upon review of the parties’ contentions and the Specification, as well
as this entire record, we also discern no reason to modify our claim
constructions at this juncture, other than the clarifications we have provided
above.
B. Principles of Law
To prevail in its challenges to the patentability of the claims,
Petitioner must prove unpatentability by a preponderance of the evidence.
35 U.S.C. § 316(e); 37 C.F.R. § 42.1(d). A claim is unpatentable under
35 U.S.C. § 102 if a single prior art reference expressly or inherently
describes each and every limitation set forth in the claim. See Perricone v.
Medicis Pharm. Corp., 432 F.3d 1368, 1375 (Fed. Cir. 2005); Verdegaal
Bros., Inc. v. Union Oil Co. of Cal, 814 F.2d 628, 631 (Fed. Cir. 1987).
IPR2015-00345
Patent 8,205,237 B2

11

Additionally, a patent claim is unpatentable under 35 U.S.C. § 103(a)
if the differences between the claimed subject matter 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. KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398,
406 (2007). 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) objective evidence
of nonobviousness. Graham v. John Deere Co., 383 U.S. 1, 17–18 (1966).
In that regard, an obviousness analysis “need not seek out precise teachings
directed to the specific subject matter of the challenged claim, for a court
can take account of the inferences and creative steps that a person of
ordinary skill in the art would employ.” KSR, 550 U.S. at 418; Translogic,
504 F.3d at 1259. We analyze the asserted grounds of unpatentability in
accordance with the above-stated principles.
C. Anticipation by Iwamura - Claims 1, 3–5, 7–9, 11–13, 15, 16, 21–25,
29, 30, 33, 37, and 38
Petitioner argues that claims 1, 3–5, 7–9, 11–13, 15, 16, 21–25, 29,
30, 33, 37, and 38 are anticipated by Iwamura. Pet. 7–16. Patent Owner
disputes this anticipation, arguing that Iwamura does not conduct a sub-
linear time search, an approximate nearest neighbor search or a non-
exhaustive search, per the independent claims. PO Resp. 11–41. Petitioner
counters these arguments. Reply 6–20. As discussed below, we determine
that Petitioner has not shown, by a preponderance of the evidence, that
IPR2015-00345
Patent 8,205,237 B2

12

claims 1, 3–5, 7, 8, 21, 22, 25, 29, 30, 33, 37, and 38 of the ’237 Patent are
anticipated by Iwamura. In contrast, we determine that Petitioner has
shown, by a preponderance of the evidence, that claims 9, 11–13, 15, 16, 23,
and 24 of the ’237 Patent are anticipated by Iwamura.
i) Petitioner’s assertions with respect to Iwamura
Iwamura discloses a system for identifying a melody input by a user,
whereby the system searches a remote music database for the melody. Ex.
1012, Abstract. After the search, the database server sends back the results
in the form of a web page. Id. Iwamura discloses the extraction of features
from an electronic work, with the melody input through a microphone,
converted into an electronic signal, and analyzed by a Fast Fourier
Transform (“FFT”) to obtain frequency spectrum information from the
waveform data. Id. at 4:4–14. The extracted features are used to search the
remote music database “using a peak or differential matching algorithm” (id.
at 12:1–2), although other matching algorithms may be applied, including a
Boyer-Moore algorithm. Id. at 9:57–10:3. The web page that the database
server sends back can include a link to an on-line music shop, which can
enable a user to make a purchase, or can include a sound file to be played by
the PC client. Id. at 12:9–13.
In addition, Iwamura discloses that “[t]o accelerate the search,
computation of the total absolute difference can be stopped when it exceeds
a certain limit.” Id. at 7:56–57. Petitioner identifies this as an example of
Iwamura’s non-exhaustive searching, in that not all records in the remote
music database are necessarily searched. Pet. 10. Also, Iwamura provides
for input fault tolerance so that the “search engine will find the closest
melody from the database” (Ex. 1012, 9:24–25), where no initial exact result
IPR2015-00345
Patent 8,205,237 B2

13

is obtained, which Petitioner argues is equivalent to the claimed
“identify[ing] a neighbor,” “identify[ing] a near neighbor,” and “to perform
an approximate nearest neighbor search.” Pet. 10–16.
ii) Claims 1, 3–5, 7, 8, 21, 22, 33, 37, and 38
Independent claims 1, 5, and 33 all recite, in part, the identification of
the media work through the use of a sub-linear or sublinear search. Patent
Owner argues that Iwamura does not teach an algorithm that “scales with a
less than linear relationship to the size of the data set to be searched” where
the data set is either the number of records in the database or “even the
length of an individual record.” PO Resp. 11. Patent Owner points out that
the Petition relies solely on the disclosure of the “Boyer-Moore algorithm”
in Iwamura, which is not sub-linear based on the claim construction adopted
above. Id. at 12 (citing Ex. 2005 ¶¶ 113–118).
Additionally, Patent Owner argues that Petitioner’s declarant, Dr.
Moulin, has acknowledged that sublinear is not “in relation with the size of
the query,” but rather in relationship to database size, and that the Boyer-
Moore algorithm is linear, as opposed to sublinear. Id. at 12–14 (citing
Ex. 2006, 24:1–12, 25:4–12, 26:11–21, 28:4–16, 44:20–46:6, 59:6–9, 61:18–
62:9, 68:25–69:4, 77:14–24). Further, Dr. Moulin acknowledged that his
representation in his Declaration was wrong. Id. at 14–16 (citing Ex. 1004
¶ 72; Ex. 2006, 67:17–21, 74:20–24, 79:9–18).
In response, Petitioner argues that Iwamura discloses a sublinear
search because “Iwamura's search speed scales at a less than linear
relationship to disk space when higher resolution works are added to a
reference database of lower resolution works.” Reply 16. This
interpretation is based on Petitioner’s view of the claim construction of “sub-
IPR2015-00345
Patent 8,205,237 B2

14

linear,” which we do not find to be persuasive and we have not adopted. See
Section II.A.i. Even assuming that songs in Imamura would occupy
different amounts of disk space, but the same number of notes, we are
persuaded that the number of records, i.e., melodies, would remain the same
and each would be compared in the search process in Imamura. See Ex.
1012, 7:53–55; Ex. 2005 ¶110. As such, we are not persuaded that Iwamura
discloses a sublinear search process; Iwamura, therefore, cannot anticipate
claims 1, 5, and 33, which recite such a search aspect.
Based on the analysis of independent claims 1, 5, and 33, we are also
persuaded that Iwamura fails to anticipate claims 3, 4, 7, 8, 21, 22, 37, and
38, by virtue of their dependence on claims 1, 5, and 33.
iii) Claims 9–11, 13–15, 23 and 24
Based on Petitioner’s assertions with respect to Iwamura, discussed
above, we are persuaded that Petitioner has demonstrated that independent
claims 9 and 13 are anticipated by Iwamura. Iwamura discloses that
received features, extracted from a media work, are identified through an
approximate nearest neighbor search of extracted features of identified
media works, with information about the identified media work transmitted
back. We discuss Patent Owner’s arguments against anticipation below.
Independent claims 9 and 13 both recite, in part, the identification of
the media work through the use of an approximate nearest neighbor search
of extracted features of identified media works. Patent Owner argues that
Iwamura does not disclose an approximate nearest neighbor search because
Iwamura does not disclose “identifying a close match that is not necessarily
the closest match.” PO Resp. 19. Patent Owner asserts that because
Iwamura discloses a search that always identifies an exact or closest match,
IPR2015-00345
Patent 8,205,237 B2

15

it cannot disclose an approximate nearest neighbor search. Id. Additionally,
Patent Owner argues that the Petition fails to explain why the fault tolerance
capability and skipped portion are relevant to or disclose an approximate
nearest neighbor search, or how they enable a search to return a result other
than the closest match. Id. at 19–23. We do not agree.
As Petitioner demonstrates, the search in Iwamura through “peak
notes” does not necessarily consider the closest match, or identify it. Pet. 8,
10, 15; Reply 13–14; Ex. 1012, 7: 37–39, 52–55, 9:36–52, 12:6–9; Ex. 2005
¶¶ 162–63. As such, if the closest matching melody segment does not occur
at a peak alignment, Iwamura will not consider or locate that segment.
Therefore, Iwamura will not always locate the closest matching melody
segment, and discloses an approximate nearest neighbor search.
Patent Owner also argues that an approximate nearest neighbor search
is a sub-linear search, and that “there is no evidence that the referenced
‘fault tolerance capability of Iwamura’ teaches a sublinear search.” PO
Resp. 25 (Citing Ex. 2005 ¶ 157), id. at 20–23. Given the discussion above
that we are not persuaded that an approximate nearest neighbor search must
be a sub-linear search, we likewise do not find this argument to be
persuasive.
With respect to dependent claims 11, 12, 15, 16, 23, and 24, we are
persuaded that Petitioner has demonstrated that Iwamura teaches the
elements of those dependent claims, as set forth in the Petition. See Pet. 11,
13–14.
iv) Claims 25, 29, and 30
Independent claim 25 recites, in part, the identification of the media
work through the use of a nonexhaustive search. Patent Owner argues that
IPR2015-00345
Patent 8,205,237 B2

16

Iwamura fails to disclose a nonexhaustive search. PO Resp. 25–38. Patent
Owner asserts that Iwamura’s peak search algorithm is designed to be more
efficient than alternative approaches, such as one in which the entered
melody is shifted “note by note” relative to a single reference melody. Id. at
25–26 (citing Ex. 1012, 12:1–2; Ex. 2005 ¶¶ 159–166; see also ¶ 312).
According to Patent Owner, Iwamura’s peak note search does not reduce the
number of reference melodies (i.e., records in the database) to be searched,
but instead speeds up the individual comparison of the work to be identified
with each record, by shifting the melody to align peak notes rather than
shifting note by note. Id. (citing Ex. 2005 ¶ 163). Each record in the
database is searched, and the reference melody that gives the least total
absolute difference when compared with the entered melody is returned as a
search result. Id. at 26 (citing Ex. 1012, 7:53–55; Ex. 2005 ¶ 163).
Accordingly, Patent Owner continues, Iwamura discloses an exhaustive
search because it searches all records in the database. Id. (citing Ex. 2005
¶ 164).
Patent Owner cites testimony from Petitioner’s declarant, Dr. Moulin,
in which he agrees that Iwamura’s system compares a work to be identified
with each reference work in the database. PO Resp. 26–28 (citing Ex. 2006,
207:18–23, 223:2–8, 247:18–20, 269:19–270:2, 271:19–21); see also
Tr. 17:17 (Petitioner conceding at oral hearing that “every song is looked
at”). Dr. Moulin also acknowledges that if a non-exhaustive search is one
that does not look at each of the melodies in the database, Iwamura discloses
an exhaustive search rather than a non-exhaustive search. Ex. 2006, 233:24–
234:14; see PO Resp. 27.
IPR2015-00345
Patent 8,205,237 B2

17

Despite Dr. Moulin’s testimony and Iwamura’s description of a search
that compares an input melody to each melody in the database, Petitioner
contends that Iwamura discloses a non-exhaustive search. First, Petitioner
submits that Iwamura specifically identifies its peak note search as a non-
exhaustive search because the “search speed using peak notes is 20% of a
brute force search.” Pet. 9–10 (quoting Ex. 1012, 9:8–11); Reply 7–8 (citing
Ex. 1012, 9:7–11; Ex. 2005 ¶ 82; Ex. 2001). We agree with Patent Owner,
however, that Iwamura’s peak note search accelerates a search within a
single comparison between a work to be identified and an individual record
in the database, while still comparing the work to each record in the
database. See PO Resp. 29–30. In other words, the peak search technique
may improve the efficiency of an individual comparison between an entered
melody and a reference melody in the database, but it is still an exhaustive
search because it compares the entered melody with each musical work in
the database, as confirmed by Petitioner’s declarant, Dr. Moulin. Id. (citing
Ex. 2006, 269:19–270:2).
Petitioner further argues that Iwamura discloses a non-exhaustive
search because the search can be accelerated by stopping the search when
computations “exceed[] a certain limit.” Pet. 10 (quoting Ex. 1012, 7:56–
57). We agree with Patent Owner that this disclosure also refers to
improving the speed of a comparison between the work to be identified and
a single record in the reference database. See PO Resp. 32 (citing Ex. 2005
¶ 179). As Dr. Karypis explains and Dr. Moulin confirms, when the total
absolute difference computed between the entered melody and a particular
reference melody exceeds a certain limit, the computation stops, and the
search shifts to the next peak range comparison within the same record.
IPR2015-00345
Patent 8,205,237 B2

18

Id. at 32–35 (citing Ex. 2005 ¶¶ 180–81; Ex. 2006, 241:24–242:2). Thus,
Iwamura’s computation “limit” does not describe a non-exhaustive search
because Iwamura still compares a work to be identified with each reference
melody in the database.
In addition, Petitioner argues that Iwamura discloses non-exhaustive
searching because its search skips “portions that should not be searched,”
such as “repeated patterns” and “unimportant portion[s].” Pet. 10 (quoting
Ex. 1012, 12:6–7, 9:36–45); Reply 13. Again, we agree with Patent Owner
that the cited passages refer to techniques for accelerating a search between
a work to be identified and a record in the database. PO Resp. 35 (citing
Ex. 2005 ¶ 185). As confirmed by Dr. Moulin, the search still compares the
work with each musical work in the reference database. Id. at 35–36 (citing
Ex. 2006, 267:13–24, 317:2–12); see also Ex. 2005 ¶ 187.
Finally, Petitioner contends that Iwamura does not consider all
“possible matches”—and therefore does not disclose a “non-exhaustive
search”—because the possible matches in Iwamura are melody segments
rather than full songs, and Iwamura does not consider all melody segments.
Reply 12–13. We are not persuaded by this argument. Iwamura explicitly
states that its search returns a reference melody as a search result. Ex. 1012,
7:54–55; see also id. at 1:53–56 (“The invented music search allows a user
to search a database and thereby obtain the title of the work only with its
melody as input to a search engine . . . .”). Furthermore, Petitioner’s
declarant, Dr. Moulin, testifies that the possible matches in Iwamura’s
search are the reference melodies in the database. Ex. 2006, 206:12–15,
217:15–18, 247:18–20. Although Petitioner cites deposition testimony of
Patent Owner’s declarant, Dr. Karypis, in which he refers to individual
IPR2015-00345
Patent 8,205,237 B2

19

comparisons of an entered melody segment to segments in a reference
melody as “matches,” Reply 12–13 (citing Ex. 1020, 134:5–135:10), he does
not describe the melody segments in the database as “possible matches.”
See Tr. 31:19–32:2. For these reasons, we are not persuaded that the
possible matches for purposes of determining whether Iwamura’s search is
non-exhaustive are melody segments. As discussed, because Iwamura’s
search compares a work with every record in the reference database, the
search is exhaustive rather than non-exhaustive.
Based on the analysis of independent claim 25, we are also persuaded
that Iwamura fails to anticipate claims 29 and 30, by virtue of their
dependence on claim 25.
v) Conclusion
We are not persuaded that Petitioner has established by a
preponderance of the evidence that claims 1, 3–5, 7, 8, 21, 22, 25, 29, 30, 33,
37, and 38 of the ’237 Patent are unpatentable under 35 U.S.C. § 102 as
being anticipated by Iwamura. However, we are persuaded that Petitioner
has established by a preponderance of the evidence that claims 9, 11–13, 15,
16, 23, and 24 of the ’237 Patent are unpatentable under 35 U.S.C. § 102 as
being anticipated by Iwamura.

D. Anticipation by Ghias - Claims 1–3, 5–7, 9–11, 13–15, and 21–24
Petitioner argues that claims 1–3, 5–7, 9–11, 13–15, and 21–24 are
anticipated by Ghias. Pet. 39–47. Patent Owner disputes this anticipation,
arguing that Ghias does not conduct a sub-linear time search or an
approximate nearest neighbor search, per the independent claims. PO Resp.
IPR2015-00345
Patent 8,205,237 B2

20

41–55. Petitioner counters these arguments. Reply 20–24. As discussed
below, we determine that Petitioner has not shown, by a preponderance of
the evidence, that claims 1–3, 5–7, 21, and 22 of the ’237 Patent are
anticipated by Ghias. In contrast, we determine that Petitioner has shown,
by a preponderance of the evidence, that claims 9–11, 13–15, 23, and 24 of
the ’237 Patent are anticipated by Ghias.
i) Petitioner’s assertions with respect to Ghias
Ghias relates to searching for melodies. Ex. 1010, Abstract. The
system of Ghias receives a melody input through a microphone, converts it
into a digitized representation based on relative pitch differences between
successive notes, and searches a database of such representations for an
approximate match. Id. Ghias also provides that a preselected error
tolerance may be applied to the search. Id. at 2:50–59. The results of the
search are presented as a ranked list of approximately matching melodies, or
alternately just one best match. Id. at 2:50–59, 6:60–63. Ghias also
discloses that it is desirable to perform key-searching within the database
using “an efficient approximate pattern matching algorithm,” where different
algorithms have various running times dependent on the number of entries in
the database, including search algorithms that are substantially faster than
“brute force” searches, with execution times that are proportional to the
logarithm of the size of the data set. Id. at 6:7–11, 6:23–35.
Ghias also discloses that “a cubic-spline wavelet transform or other
suitable wavelet transform” may be used to improve the speed and
robustness of the pitch-tracking algorithm. Id. at 5:13–20. As well, Ghias
provides that the melody database “may be located apart from computer 16
and suitably connected thereto for communicating.” Id. at 6:13–20.
IPR2015-00345
Patent 8,205,237 B2

21

ii) Claims 1–3, 5–7, 21, and 22
Independent claims 1 and 5 recite, in part, the identification of the
media work through the use of a sub-linear or sublinear search. Patent
Owner argues that Ghias teaches a linear search in which the search time
grows linearly in relationship to the size of the data set. PO Resp. 42 (citing
Ex. 2005 ¶¶ 213–233). Patent Owner continues that Ghias provides that the
user input is compared with all songs, such that the execution time of the
search would increase linearly with the size of the data set. Id. (citing Ex.
1010, 5:66-6:2; Ex. 2005 ¶ 213).
We agree with Patent Owner that the disclosed searches in Ghias are
linear, not sublinear, such that Ghias fails to teach at least one aspect of
independent claims 1 and 5. Although Petitioner asserts that Ghias
anticipates claims 1 and 5 (Reply 1), Petitioner offers no responsive
arguments with respect to claims 1 and 5, or claims dependent thereon. See
Reply 20–24. As such, we are not persuaded that Petitioner has
demonstrated that Ghias anticipates independent claims 1 and 5.
Based on the analysis of independent claims 1 and 5, we are also
persuaded that Iwamura fails to anticipate claims 2, 3, 6, 7, 21, and 22 by
virtue of their dependence on claims 1 and 5.
iii) Claims 9–11, 13–15, 23, and 24
Based on Petitioner’s assertions with respect to Ghias, discussed
above, we are persuaded that Petitioner has demonstrated that independent
claims 9 and 13 are anticipated by Ghias. Ghias discloses that received
features, extracted from a media work, are identified through an approximate
nearest neighbor search of extracted features of identified media works, with
IPR2015-00345
Patent 8,205,237 B2

22

information about the identified media work transmitted back. We discuss
Patent Owner’s arguments against anticipation below.
Independent claims 9 and 13 both recite, in part, the identification of
the media work through the use of an approximate nearest neighbor search
of extracted features of identified media works. Patent Owner argues that
Ghias does not disclose an approximate nearest neighbor search because
Ghias does not disclose “identifying a close match that is not necessarily the
closest match.” PO Resp. 51. Patent Owner argues that Ghias teaches that a
search generates three possible outputs, including an exact match, a ranked
list, and the single most approximate matching melody, but each of these
outputs always identifies the closest match. Id. at 51-53 (citing Ex. 2005
¶¶ 239–424; Ex. 2006, 341:23–342:1, 345:16–346:11, 352:22–353:2, 356:
2–21). We do not agree.
As Petitioner demonstrates, Ghias describes a first search, which
considers all reference works, followed by subsequent searches to refine the
results. Pet. 42 (citing Ex. 1010, 2:50–59, 6:60–63, 7:5–8; Ex. 1004 ¶ 124);
Reply 21–22. The new query for a subsequent search can be a different song
or a different part or version of the same song, such that the first search may
have excluded works that would be closest matches for the new query.
Reply 21 (citing Ex. 2005 ¶ 287; Ex. 1020, 172:5–173:3). As such, the
search process in Ghias can locate a close match, but need not always
generate an exact or closest match. Therefore, we are persuaded that
Petitioner has demonstrated that Ghias discloses an approximate nearest
neighbor search.
Patent Owner also argues that “Ghias does not disclose a sublinear
search.” PO Resp. 53–54 (citing Ex. 2005 ¶ 246). Given the discussion
IPR2015-00345
Patent 8,205,237 B2

23

above that we are not persuaded that an approximate nearest neighbor search
must be a sub-linear search, we likewise do not find this argument to be
persuasive.
With respect to dependent claims 10, 11, 14, 15, 23, and 24, we are
persuaded that Petitioner has demonstrated that Ghias teaches the elements
of those dependent claims, as set forth in the Petition. See Pet. 43–47.
iv) Conclusion
We are not persuaded that Petitioner has established by a
preponderance of the evidence that claims 1–3, 5–7, 21, and 22 of the ’237
Patent are unpatentable under 35 U.S.C. § 102 as being anticipated by Ghias.
However, we are persuaded that Petitioner has established by a
preponderance of the evidence that claims 9–11, 13–15, 23, and 24 of the
’237 Patent are unpatentable under 35 U.S.C. § 102 as being anticipated by
Ghias.

E. Obviousness Over Iwamura and Chen - Claims 26, 27, 34, and 35
Petitioner argues that claims 26, 27, 34, and 35 would have been
obvious over Iwamura in combination with Chen. Pet. 53–57. This ground
is based on the ground of claims 1, 3–5, 7–9, 11–13, 15, 16, 21–25, 29, 30,
33, 37, and 38 being anticipated by Iwamura, discussed above. Pet. 53.
Patent Owner argues that Iwamura does not disclose a non-exhaustive search
to identify a near neighbor or a sublinear approximate nearest neighbor
search, as recited in independent claims 25 and 33 discussed above, and that
Petitioner does not rely on Chen for those elements. PO Resp. 56–57. We
agree.
IPR2015-00345
Patent 8,205,237 B2

24

Claims 26, 27, 34, and 35 depend from independent claims 25 and 33,
where, as discussed above, those independent claims recite either that the
search is non-exhaustive or sublinear. As we are not persuaded that
Iwamura anticipates independent claims 25 and 33, we are likewise
unpersuaded that Petitioner has demonstrated that claims 26, 27, 34, and 35
are unpatentable over Iwamura and Chen, with the ground relying on
Iwamura to teach or suggest the non-exhaustive or sublinear searching
recited in the independent claims.

III. CONCLUSION
For the foregoing reasons, we conclude that Petitioner has
demonstrated, by a preponderance of the evidence, that claims 9–16, 23, and
24 of the ’237 Patent are unpatentable as being anticipated by Iwamura
under 35 U.S.C. § 102(e), and that Petitioner has not demonstrated, by a
preponderance of the evidence, that claims 1–8, 21, 22, 25–27, 29, 30, 33–
35, 37, and 38 of the ’237 Patent are unpatentable.

IV. ORDER
In consideration of the foregoing, it is
ORDERED that claims 9–16, 23, and 24 of the ’237 Patent have been
shown to be unpatentable;
FURTHER ORDERED that claims 1–8, 21, 22, 25–27, 29, 30, 33–35,
37, and 38 of the ’237 Patent have not been shown to be unpatentable; and
IPR2015-00345
Patent 8,205,237 B2

25

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.

IPR2015-00345
Patent 8,205,237 B2

26

For PETITIONER:
James J. Elacqua
Douglas R. Nemec
SKADDEN, ARPS, SLATE, MEAGHER & FLOM
DLGOOGLE1N@skadden.com
James.Elacqua@skadden.com
Douglas.Nemec@skadden.com

For PATENT OWNER:
Charles R. Macedo
Brian A. Comack
AMSTER, ROTHSTEIN & EBENSTEIN LLP
cmacedo@arelaw.com
N1-Google-IPR@arelaw.com
Gregory Dovel
DOVEL & LUNER, LLP
greg@dovellaw.com