`
`
`On behalf of Amazon.com, Inc.
`By:
`Joseph R. Re
`
`Colin B. Heideman
`
`Christie R.W. Matthaei
`
`KNOBBE, MARTENS, OLSON & BEAR, LLP
`925 Fourth Avenue, Suite 2500
`Seattle, WA 98104
`Telephone: (206) 405-2000
`Facsimile: (206) 405-2001
`Email: BoxSEAZNL1495LP3@knobbe.com
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`__________________________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`__________________________________
`
`AMAZON.COM, INC.,
`Petitioner,
`
`v.
`
`CUSTOMPLAY, LLC,
`Patent Owner.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`IPR2018-01498
`Patent 9,380,282
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`PETITIONER’S REPLY
`
`
`
`TABLE OF CONTENTS
`
`ARGUMENT ----------------------------------------------------------------------------- 1
`
`I.
`
`GROUND 1A: OBVIOUSNESS OVER MCINTIRE AND DEY --------- 2
`
`A.
`
`Claims 4, 9, 12, 14, and 19 ---------------------------------------------- 2
`
`1.
`
`Contemporaneously Retrieving a Second VFI ---------------- 2
`
`a. McIntire Discloses This Limitation --------------------- 2
`
`b.
`
`c.
`
`Obviousness Over McIntire ------------------------------ 4
`
`Obviousness Over McIntire and Dey ------------------- 5
`
`2.
`
`The Contemporaneously Displaying Limitation -------------- 5
`
`a. McIntire Discloses This Limitation --------------------- 5
`
`b.
`
`c.
`
`d.
`
`Obviousness Over McIntire ------------------------------ 6
`
`Dey Discloses This Limitation --------------------------- 7
`
`Obviousness Over McIntire & Dey --------------------- 9
`
`i.
`
`ii.
`
`iii.
`
`McIntire Does Not Teach Away ---------------- 10
`
`Dey Does Not Teach Away ---------------------- 11
`
`The Combination Does Not Change
`McIntire’s Mode of Operation ------------------ 11
`
`B.
`
`Claim 16: Additional Information ------------------------------------- 12
`
`II. GROUND 1B: MCINTIRE, DEY, & ABECASSIS ------------------------ 13
`
`A.
`
`Claim 7 -------------------------------------------------------------------- 13
`
`1.
`
`2.
`
`Obviousness Over McIntire ------------------------------------- 13
`
`Obviousness Over Abecassis ----------------------------------- 14
`
`-i-
`
`
`
`B.
`
`Claims 8 and 18 ---------------------------------------------------------- 15
`
`III. GROUND 2A: BERGEN & REIMER ---------------------------------------- 15
`
`A.
`
`Bergen Discloses Identifying the Request Location ----------------- 15
`
`1. Motivation to Combine Bergen & Reimer -------------------- 17
`
`2.
`
`The Combination Does Not Change Bergen’s
`Principle of Operation ------------------------------------------- 17
`
`B.
`
`C.
`
`Bergen Discloses the Retrieving Limitations ------------------------- 18
`
`Bergen Discloses the Displaying Limitations ------------------------ 18
`
`IV. GROUND 2B: BERGEN, REIMER, & ABECASSIS --------------------- 19
`
`A.
`
`B.
`
`Claim 7 -------------------------------------------------------------------- 19
`
`Claim 8 -------------------------------------------------------------------- 20
`
`V. GROUND 3A: ARMSTRONG ------------------------------------------------ 20
`
`A.
`
`Claim 4 -------------------------------------------------------------------- 20
`
`1.
`
`2.
`
`3.
`
`4.
`
`Armstrong Discloses Identifying a Request Location ------- 20
`
`Armstrong Discloses Retrieving a First VFI ------------------ 21
`
`Armstrong Discloses Contemporaneously Retrieving
`a Second VFI ------------------------------------------------------ 22
`
`Armstrong Discloses the Displaying Limitations ------------ 23
`
`a.
`
`b.
`
`c.
`
`Armstrong’s Background Images and Menus
`Are “Information” ---------------------------------------- 23
`
`The Displayed “Information” Need Not Be the
`Same Type ------------------------------------------------- 24
`
`Petitioner’s Mapping of Armstrong in the ’950
`IPR Is Irrelevant ------------------------------------------ 24
`
`-ii-
`
`
`
`B.
`
`C.
`
`Claim 9 -------------------------------------------------------------------- 25
`
`Claim 16 ------------------------------------------------------------------- 25
`
`VI. GROUND 3B: ARMSTRONG & ABECASSIS ---------------------------- 26
`
`CONCLUSION -------------------------------------------------------------------------- 27
`
`
`
`
`
`-iii-
`
`
`
`TABLE OF AUTHORITIES
`
`Cases:
`
`Page(s):
`
`Allied Erecting & Dismantling Co. v. Genesis Attachments, LLC,
`825 F.3d 1373 (Fed. Cir. 2016) ------------------------------------------- 12, 17
`
`Bayer Pharma AG v. Watson Labs., Inc.,
`874 F.3d 1316 (Fed. Cir. 2017) ------------------------------------------------ 10
`
`CRFD Research, Inc. v. Matal,
`876 F.3d 1330 (Fed. Cir. 2017) ------------------------------------------------ 10
`
`DePuy Spine, Inc. v. Medtronic Sofamor Danek, Inc.,
`567 F.3d 1314 (Fed. Cir. 2009) ------------------------------------------------ 10
`
`KSR Int’l Co. v. Teleflex Inc.,
`550 U.S. 398 (2007) ------------------------------------------------------------- 13
`
`In re NuVasive, Inc.,
`842 F.3d 1376 (Fed. Cir. 2016) ------------------------------------------------- 4
`
`Samsung Elecs. Co., Ltd. v. ProMOS Techs, Inc.,
`IPR2017-01413, Paper 33 (P.T.A.B. October 22, 2018) -------------------- 4
`
`Smith & Nephew, Inc. v. Rea,
`721 F.3d 1371 (Fed. Cir. 2013) ------------------------------------------------ 15
`
`Miscellaneous:
`
`Trial Practice Guide Update (August 2018) --------------------------------------- 4
`
`
`
`
`
`-iv-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`ARGUMENT
`
`It is undisputed that receiving a user request for information during the
`
`playing of a video, identifying the request location, retrieving a video frame
`
`identifier (“VFI”) responsive to the request location, and displaying information
`
`responsive to that VFI was known in the art. (See Ex. 1101, 7:3-8:15, 21:4-22:21,
`
`34:1-35:16, 45:20-46:22, 91:5-12, 110:21-111:15.) Patent Owner (“PO”) argues
`
`that the challenged claims are patentable because they purport to solve a user-delay
`
`problem by reciting the contemporaneous retrieval of two VFIs and the
`
`contemporaneous display of information associated with each of those VFIs.
`
`(POR, 5-6). But even PO’s expert (“Reader”) admits that the user-delay problem
`
`was known in the art (Ex. 1101, 154:13-16) and that the solution claimed in ’282
`
`patent was both known (id. 157:10-158:5, 160:4-14) and within a POSITA’s skill
`
`(id., 45:14-19, 49:2-55:12, 116:16-20, 122:13-17, 126:11-14, 129:13-18).
`
`Rather than advancing the art, the ’282 patent removes from the public
`
`systems and methods of displaying information that were already known and that
`
`PO did not invent. The arguments in the PO Response (“POR”) ignore key
`
`disclosures in the prior art, ignore the knowledge of a POSITA, and are
`
`contradicted by Reader’s cross-examination testimony.
`
`The challenged claims should be canceled.
`
`-1-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`I. GROUND 1A: OBVIOUSNESS OVER MCINTIRE AND DEY
`
`PO begins by mischaracterizing McIntire and arguing it teaches away from
`
`“contemporaneously displaying.” (POR, 13-20.) As discussed below, PO is
`
`wrong.
`
`A. Claims 4, 9, 12, 14, and 19
`
`PO does not dispute, and Reader admitted (Ex. 1101, 163:1-21), that
`
`McIntire discloses the “receiving” and “identifying” steps. PO admits McIntire
`
`discloses retrieving a first VFI. (POR, 15.)
`
`1.
`
`Contemporaneously Retrieving a Second VFI
`a. McIntire Discloses This Limitation.
`
`PO does not dispute that McIntire discloses contemporaneously retrieving
`
`multiple VFIs. (POR, 20-21; Ex. 2021 ¶¶78-79.) Rather, PO incorrectly argues
`
`the VFIs are responsive to a single request location because they are in a single
`
`segment. (Id.) That argument ignores the relevant disclosure in McIntire and
`
`contradicts both experts’ testimony.
`
`First, McIntire does not disclose only retrieving supplemental content for a
`
`single segment. As Reader admitted, “McIntire discloses retrieving supplemental
`
`content responsive to multiple segment identifiers at the same time.” (Ex. 1101,
`
`164:3-6; id., 163:22-164:2 (McIntire retrieves content for “multiple segments at a
`
`time”), 161:11-14 (McIntire retrieves content “by first identifying the segment
`
`-2-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`identifier”).) Thus, McIntire discloses retrieving multiple segment identifiers—not
`
`just multiple frames within a single segment—and then retrieving the supplemental
`
`content associated therewith. (Ex. 1004 ¶[0311] (server “matches one or more
`
`segment[] identifiers identifying segments that correspond to viewer signals to the
`
`appropriate supplemental content”), ¶[0312] (articles appear in an associated
`
`“portion”—not single segment—of media); Ex. 1002 ¶¶55-56; Ex. 1102 ¶¶6, 12.)
`
`Second, PO ignores that McIntire discloses an embodiment in which a
`
`“segment” is a single frame, and this is the embodiment Petitioner relies on. (Pet.,
`
`17-25.) Reader admitted that, in such an embodiment, McIntire’s “segment
`
`identifier” is a “video frame identifier.” (Ex. 1101, 160:21-161:21.) Thus,
`
`Reader’s admission that McIntire discloses retrieving information responsive to
`
`“multiple segment identifiers” (id. 164:3-6) is an admission that McIntire discloses
`
`retrieving two different VFIs in different “segments.” And, a POSITA would have
`
`understood that the second VFI would be “prior to the request location.” (Ex. 1002
`
`¶¶54-56, 74; Ex. 1102 ¶¶8, 13-15.) Neither PO nor Reader disputes that.
`
`Third, PO ignores—but Reader admitted—that McIntire discloses retrieving
`
`multiple segment (video frame) identifiers “around” the time of the user’s request,
`
`and retrieving information associated with those identifiers. (Pet., 17-22; Ex.
`
`1101, 164:7-12.) Where McIntire’s segments are individual frames, McIntire’s
`
`system retrieves, from a plurality of VFIs (all VFIs for the video), a first VFI (first
`
`-3-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`segment identifier for current frame) that is responsive to the request location, and
`
`contemporaneously retrieves a second, different VFI that is responsive to a prior
`
`location (a second VFI that identifies a single prior frame “around” the time of the
`
`request). (Ex. 1102 ¶¶14-15.)
`
`PO never disputes, and Reader’s testimony confirms, that McIntire discloses
`
`this limitation when McIntire’s segments are individual frames. PO waived any
`
`such argument. In re NuVasive, Inc., 842 F.3d 1376, 1380-81 (Fed. Cir. 2016)
`
`(argument not raised in POR is waived); Samsung Elecs. Co., Ltd. v. ProMOS
`
`Techs, Inc., IPR2017-01413, Paper 33 at 20 (P.T.A.B. October 22, 2018) (refusing
`
`to consider new argument in PO’s sur-reply); Trial Practice Guide Update (August
`
`2018) at 15 (sur-reply may not raise new issues, belatedly present evidence, or
`
`“embark in a new direction”); Scheduling Order (Paper 14) at B.1.a (“any
`
`arguments not raised in the response may be deemed waived”).
`
`b. Obviousness Over McIntire
`
`PO offers a single conclusory sentence that McIntire does not teach or
`
`suggest this limitation. (POR, 21.) But PO never rebuts Petitioner’s arguments
`
`(Pet., 19-20) or Bovik’s testimony (Ex. 1002 ¶¶56-59).
`
`Reader’s testimony supports Petitioner’s argument. Reader confirmed that
`
`the user-delay problem was known in the art. (Ex. 1101, 154:13-19.) And both
`
`experts agree that: (1) because of the known user-delay problem, a POSITA would
`
`-4-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`have been motivated to provide supplemental content associated with frames prior
`
`to the request location (id., 154:20-155:14, 157:11-158:5; Ex. 1002 ¶74); (2) a
`
`POSITA knew they could contemporaneously retrieve information associated with
`
`the current frame and a prior frame (Ex. 1101, 160:4-14; Ex. 1002 ¶¶57-59); and
`
`(3) retrieving two VFIs responsive to two different locations was within a
`
`POSITA’s skill (Ex. 1101, 45:14-19; Ex. 1002 ¶59). Thus, this limitation would
`
`have been obvious over McIntire alone.
`
`c.
`
`Obviousness Over McIntire and Dey
`
`PO argues Dey “teaches away” from the contemporaneously retrieving step
`
`for the same reasons Dey allegedly “teaches away” from the contemporaneously
`
`displaying limitation. (POR, 19-20.) That attorney argument is unsupported by
`
`Reader (Ex. 2021 ¶¶77-81) and is wrong, as discussed immediately below.
`
`2.
`
`The Contemporaneously Displaying Limitation
`a. McIntire Discloses This Limitation.
`
`PO argues that, when McIntire’s two frame identifiers are in the same
`
`segment, the retrieved information is the same. (POR, 21-22.) But that is not the
`
`McIntire embodiment Petitioner relies on.
`
`Both experts agree that McIntire discloses retrieving supplemental content
`
`associated with multiple segment identifiers, and that McIntire’s “segment” can be
`
`an individual frame. (See Section I.A.1.a; Pet., 17, 28.) Accordingly, contrary to
`
`-5-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`PO’s assertion, McIntire discloses retrieving two VFIs representing two different
`
`segments. (Id.) McIntire further discloses displaying the “retrieved information.”
`
`(Ex. 1004, Fig. 9; id., Fig. 8 (“display supplemental content”); Ex. 1102 ¶¶21-22;
`
`Ex. 1101, 111:8-15.) Thus, McIntire discloses this limitation.
`
`McIntire also discloses displaying different information associated with each
`
`segment (frame) identifier because the articles depicted, and therefore the
`
`associated information, would change depending on the segment (frame). (Ex.
`
`1102 ¶¶19-20.) This change causes the user-delay problem McIntire describes.
`
`(Id.; Ex. 1004 ¶[0272].) Indeed, the purpose of identifying additional frame
`
`identifiers “around” the current location and retrieving supplemental content for
`
`multiple VFIs is to display different information associated with previously
`
`depicted items that are not depicted in later locations. (Pet., 19-20; Ex. 1102 ¶20.)1
`
`b. Obviousness Over McIntire
`
`This limitation also would have been obvious over McIntire alone. Both
`
`experts agree that POSITAs knew about the user-delay problem and were
`
`motivated to solve it by providing different supplemental content associated with
`
`video frames before the request location. (Ex. 1101, 154:20-155:14, 157:11-158:5,
`
`1 McIntire discloses contemporaneously retrieving and contemporaneously
`
`displaying because each of those functions is performed in a single step. (Ex. 1004
`
`¶[0312]; Figs. 8-9; Ex. 1102 ¶¶15, 21.)
`
`-6-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`49:7-11; Ex. 1002 ¶57-59, 74; Ex. 1102 ¶23.) It is undisputed that doing so was
`
`within a POSITA’s skill. (Ex. 1101, 49:7-11, 51:6-10, 52:16-55:12.)
`
`c.
`
`Dey Discloses This Limitation.
`
`PO relies on the Board’s preliminary finding that Dey “seem[s] to teach”
`
`displaying information associated with a prior portion of the video only. (POR,
`
`22-23.) That preliminary finding was incorrect.
`
`First, while Dey discloses embodiments that do not retrieve content for “the
`
`instant” interest is expressed (Ex. 1023, 7:37-41), Dey discloses that the “more
`
`realistic model” assumes that there is “some but lesser interest” in material
`
`between t1 and the request time (id., 7:59-66). Indeed, Dey repeatedly discloses
`
`that interest may be non-zero at the request time. (See id., 4:60-65 (“interest may
`
`decrease from t1 until the time of the signal … and may be zero for all other
`
`times”), 5:2-4 (“W(t) may equal 0 for all times earlier than 30 seconds before the
`
`signal of interest is given, and later than the signal of interest.”); Fig. 5 (showing
`
`interest at time of request).) Thus, Dey discloses retrieving and displaying
`
`information for the request time and for times prior to the request time. (Ex. 1102
`
`¶24.)
`
`Second, Dey discloses providing supplemental content associated with the
`
`request location because Dey discloses that significant interest exists at t1, which
`
`may be just 0.0001 milliseconds before the request (Ex. 1023, 8:30-35, 9:56-62),
`
`-7-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`and therefore that “some” interest exists in material within 0.0001 milliseconds of
`
`the request. In the video context, a frame commonly occupies 33 milliseconds.
`
`(Ex. 1102 ¶25.) Thus, retrieving information for 0.0001 milliseconds (or less)
`
`prior to the request is retrieving information associated with the request location.
`
`(Id.; Ex. 1101, 101:6-16.)
`
`Third, even if Dey disclosed only retrieving information associated with
`
`times prior to the request location (e.g., t1 and t2), it would still disclose this
`
`limitation. As Reader admitted, the frame at t1 is a VFI “responsive to the request
`
`location” because a VFI can be “responsive to” a request location even if it is
`
`“offset” from the request location. (Ex. 1101, 100:14-101:16, 104:11-17; see Ex.
`
`1001, 14:14-18.) Thus, Dey discloses retrieving a first VFI (the frame at t1)
`
`responsive to the request location and a second prior VFI (the frame at t2), and
`
`displaying different information associated with those VFIs. (Ex. 1023, 7:45-8:9.)
`
`Finally, Dey provides that the time of interest may vary based on
`
`considerations which will be apparent to a POSITA. (Ex. 1023, 9:50-62.) Thus, a
`
`POSITA providing supplemental content as described in McIntire would have
`
`understood that the supplemental content should include content mapped to the
`
`current frame and prior frames. (Ex. 1102 ¶26.)
`
`
`
`-8-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`d. Obviousness Over McIntire & Dey
`
`The POR and Institution Decision suggest this limitation could not have
`
`been obvious unless McIntire or Dey discloses contemporaneously displaying
`
`information from the request location and a prior location. That is incorrect.
`
`It is undisputed that McIntire discloses: (a) retrieving a first VFI responsive
`
`to the request location; (b) displaying information associated with that VFI; (c)
`
`retrieving a second VFI “around” the time of the user’s request; (d) retrieving
`
`supplemental content associated with the first and second VFIs; (e) the user-delay
`
`problem; and (f) a desire to solve that problem. (Pet. 28; Ex. 1002 ¶¶48-56, 74.) It
`
`is also undisputed that Dey discloses: (g) the same user-delay problem; and (h)
`
`solving that problem by retrieving content displayed up to 30 seconds before the
`
`request. (Pet., 21-22; Ex. 1002 ¶60.)
`
`Thus, while McIntire discloses providing supplemental content associated
`
`with the request location and “around” that location, Dey teaches that the most
`
`relevant material is before—not just “around”—the request location. This would
`
`have motivated a POSITA to use McIntire’s system to display information from
`
`“prior to” the request location. (Ex. 1002 ¶¶61-70, 75.) And, because McIntire’s
`
`segment is a single frame, McIntire’s system would use a second VFI (prior frame
`
`number) to retrieve supplemental content from the prior location (as taught by
`
`Dey), and to display information associated with both locations so the user receives
`
`-9-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`the information of greatest interest. (Id.) The limitation would have been obvious
`
`even if neither McIntire nor Dey individually disclose the contemporaneously
`
`displaying limitation. (Ex. 1102 ¶¶27, 29, 30.)
`
`i. McIntire Does Not Teach Away.
`
`PO suggests McIntire “teaches away” because “McIntire discloses that the
`
`display of retrieved supplemental content is responsive to a single segment
`
`identifier.” (POR, 14-16.) As discussed above, even PO’s expert agrees that is
`
`wrong. (See Section I.A.1.a.) Additionally, McIntire “does not ‘criticize,
`
`discredit, or otherwise discourage’ investigation into the invention claimed.” (Ex.
`
`1102 ¶9); DePuy Spine, Inc. v. Medtronic Sofamor Danek, Inc., 567 F.3d 1314,
`
`1327 (Fed. Cir. 2009) (citation omitted).
`
`PO also argues that McIntire’s “statements regarding preferences are
`
`relevant” to a POSITA’s motivation. (POR, 16-18.) But McIntire states no
`
`preference. There were two methods of solving the user-delay problem. (Ex. 1002
`
`¶¶59-69.) McIntire’s disclosure of one option would not preclude or dissuade a
`
`POSITA from considering the other alternative. (Ex. 1102 ¶9); Bayer Pharma AG
`
`v. Watson Labs., Inc., 874 F.3d 1316, 1327 (Fed. Cir. 2017) (preference does not
`
`teach away from a “lesser preferred but still workable option”); CRFD Research,
`
`Inc. v. Matal, 876 F.3d 1330, 1347-48 (Fed. Cir. 2017) (reference can motivate
`
`POSITA to consider other ways to satisfy goals, even where it discloses an
`
`-10-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`embodiment satisfying the goal). Moreover, Reader admitted this option would
`
`have been known to, and within the skill of, a POSITA. (Ex. 1101, 49:2-11,
`
`51:6-55:11.)
`
`PO argues, without support from Reader, that Petitioner uses hindsight to
`
`combine a step for annotating a media stream with a method for providing
`
`annotated content. (POR, 18.) But both methods are “for providing annotated
`
`media streams to viewers.” (Ex. 1004 ¶¶[0308], [0265].) And both processes are
`
`virtually identical, as both rely on mapping segment identifiers to supplemental
`
`information. (Id. ¶¶[0272], [0311]; Ex. 1102 ¶22.)
`
`ii.
`
`Dey Does Not Teach Away.
`
`PO argues Dey teaches away because Dey allegedly discloses that the user is
`
`interested in information only before the request. (POR, 19-20.) This is incorrect.
`
`(See Section I.A.2.c.) But regardless, Petitioner does not propose modifying Dey.
`
`It would have been obvious to modify McIntire, which already displays
`
`information associated with the first VFI, to include information associated with a
`
`prior VFI, which Dey teaches is of interest. (Ex. 1002 ¶75; Ex. 1102 ¶30.)
`
`iii. The Combination Does Not Change
`McIntire’s Mode of Operation.
`
`PO argues the combination would change McIntire’s principle of operation.
`
`(POR, 24.) But Petitioner proposes merely using McIntire’s system, which Reader
`
`admits retrieves and displays information from multiple VFIs, to display
`
`-11-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`information for at least one location prior to the request, given Dey’s teachings.
`
`(Ex. 1002 ¶¶60-70, 74-75; Ex. 1102 ¶30.) No change is necessary, and the
`
`systems’ physical combinability is irrelevant. (Id.); Allied Erecting & Dismantling
`
`Co. v. Genesis Attachments, LLC, 825 F.3d 1373, 1381 (Fed. Cir. 2016).
`
`B. Claim 16: Additional Information
`
`PO’s argument that McIntire and Dey do not disclose or render obvious
`
`displaying information associated with a second VFI (POR, 25-27) is wrong for the
`
`reasons discussed for claim 4.
`
`PO also argues that “the user is not provided a choice between information
`
`associated with the first VFI or second VFI.” (POR, 27.) PO is incorrect.
`
`McIntire discloses presenting supplemental content about multiple articles
`
`appearing in a “portion of a media stream.” (Ex. 1004 ¶[0313], Fig. 10.) As
`
`discussed above, one article may be associated with the current frame while
`
`another may be associated with a frame “around” (e.g., prior to) the current frame.
`
`(Ex. 1002 ¶74; Ex. 1102 ¶¶34-35.)
`
`McIntire further discloses that the user can request—and the system would
`
`enable the display of—additional information about any item shown. (Ex. 1004,
`
`Fig. 10, ¶[0313]; Ex. 1102 ¶35; Pet., 30-33.) Moreover, Reader admitted that
`
`receiving a request for additional information, retrieving that information, and
`
`displaying it was known. (Ex. 1101, 55:13-59:8.)
`
`-12-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`II. GROUND 1B: MCINTIRE, DEY, & ABECASSIS
`A. Claim 7
`
`PO does not dispute that McIntire discloses the pausing limitation. (Pet.,
`
`31-32.) Reader agreed this was well-known. (Ex. 1101, 59:14-60:8.) The only
`
`dispute is whether resuming playing at the beginning of a video clip was obvious.
`
`1. Obviousness Over McIntire
`
`PO does not dispute that resuming play at the beginning of a clip was
`
`generally known and disclosed by Hammoud as one of four options. (POR,
`
`28-29.) It is also undisputed that a POSITA knew the benefits of resuming play at
`
`the beginning of a clip. (Ex. 1101, 141:19-22, 147:6-10; Ex. 1002 ¶92; Ex. 1102
`
`¶¶41-42.) Thus, it was obvious. KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 421
`
`(2007).
`
`PO argues that the ’282 patent discloses “an extensive list of additional
`
`options.” (POR, 28.) But the patent lists the same options as Hammoud, albeit
`
`using different terminology. (Ex. 1001, 11:65-12:19; Ex. 1102 ¶¶41-42.) The
`
`patent discloses only one additional option. (Ex. 1102 ¶41.) Thus, even PO does
`
`not identify more than five options. But whether four or five options exist, it is a
`
`finite number, and the claimed option was known and within a POSITA’s skill.
`
`(Id.)
`
`-13-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`2. Obviousness Over Abecassis
`
`PO argues Abecassis discloses resuming at the beginning of a “segment,”
`
`and that a “segment” is not a “clip.” (POR, 29-30.) That is incorrect. The ’282
`
`patent defines “clip” as a “segment that is smaller than a chapter.” (Ex. 1001,
`
`3:59-65.) Abecassis likewise discloses that a “segment” is a “part of a video” (Ex.
`
`1024, 6:66-67) that is smaller than a chapter (id., 10:38-45). Moreover, both
`
`Abecassis’ “segment” and the ’282 patent’s “clip” are defined by characters,
`
`locations, etc. (Id., 8:14-20, 39:64-40:11; Ex. 1001, 3:59-65.) Thus, Abeccassis’
`
`“segment” is a “clip” as defined in the ’282 patent. (Ex. 1102 ¶¶37-38, 44-45.)
`
`Moreover, the beginning of a segment coincides with the beginning of a clip. (Id.
`
`¶38.)
`
`Even if Abecassis did not expressly disclose resuming at the beginning of a
`
`clip, doing so would have been obvious in view of Abecassis’ teachings.
`
`Abecassis provides the motivation for a POSITA to resume at the beginning of a
`
`clip. (Ex. 1024, 5:6-12 (“viewer may re-engage the video without the loss of
`
`continuity); Ex. 1102 ¶42.) And, Reader admitted that a POSITA “would have
`
`known about the benefits of resuming at the beginning of a clip.” (Ex. 1101,
`
`147:6-10; id. 137:20-138:22.)
`
`Resuming at the beginning of the current clip was a simple design choice for
`
`which the ’282 patent describes no problems. (Ex. 1102 ¶¶41-42.) PO cannot now
`
`-14-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`argue it would have been nonobvious, problematic, or outside a POSITA’s skill.
`
`Smith & Nephew, Inc. v. Rea, 721 F.3d 1371, 1381 (Fed. Cir. 2013).
`
`PO does not dispute any of Petitioner’s motivations to combine. Thus, claim
`
`7 would have been obvious in view of McIntire, Dey, and Abecassis.
`
`B. Claims 8 and 18
`
`PO’s attorney argument for claim 8 (POR, 30-31) should be rejected.
`
`Petitioner explained that a POSITA would have known that McIntire’s video need
`
`not be paused where initial information is an image shown in a separate display
`
`window, but would need to be paused where the “additional information” is a
`
`video that would distract the viewer. (Pet., 36-37.) Neither PO nor its expert
`
`dispute that. (Ex. 2021 ¶106.) And Reader confirmed Petitioner’s positions. (Ex.
`
`1101, 66:18-69:15, 59:14-64:1.)
`
`III. GROUND 2A: BERGEN & REIMER
`
`It is undisputed that Bergen’s system receives a user request during a video
`
`and displays responsive information. (Ex. 2021 ¶126; POR, 41.)
`
`A. Bergen Discloses Identifying the Request Location.
`
`PO argues Bergen’s system does not necessarily identify a request location
`
`because “an equally plausible understanding” is that it could use some
`
`unidentified, unexplained “image recognition methodology.” (POR, 33-34.) PO
`
`cites to Reader’s declaration, but Reader states that an “equally plausible
`
`-15-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`understanding” is that Bergen does not use the methods/subsystems that Bergen
`
`itself discloses. (Ex. 2021 ¶128-29; Ex. 1102 ¶¶55-56.) Reader suggests Bergen
`
`might use alternatives instead of the methods Bergen actually discloses. (Id.)
`
`Reader never disputes that the system Bergen actually discloses identifies the
`
`request location. (See id. ¶136 (addressing a hypothetical “instance” using object
`
`recognition rather than the embodiment Petitioner relies on, which involves a user
`
`selection and the identification of the request location).)2
`
`PO ignores Bergen’s disclosure of using the request location (e.g., frame
`
`where request is made) to retrieve information about an object at that location.
`
`(Pet., 42-43; Ex. 1028, 15:37-16:10, 10:13-16.) For example, Bergen discloses
`
`receiving and processing a request from a viewer for all frames containing a
`
`specific actor or object. (Ex. 1028, 14:10-57.) In response, the system generates a
`
`“multi-dimensional feature vector descriptor representation” of the frame on which
`
`the request was made and compares that to a database to retrieve matching data.
`
`(Id., 14:66-15:27; Fig. 8.) To generate the “feature vector,” the system retrieves
`
`the frame on which the request was made and inputs the frame into a processing
`
`step. (Id., 6:12-15, Figs. 9-10; Ex. 1102 ¶61.) Thus, Bergen discloses identifying
`
`a request location responsive to the request for information.
`
`2 Even if image recognition were a plausible alternative, a system using
`
`image recognition would still identify the request location. (Ex. 1102 ¶56.)
`
`-16-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`1. Motivation to Combine Bergen & Reimer
`
`PO does not dispute that Reimer discloses this limitation. PO argues only
`
`that no motivation to combine exists. (POR, 35-36.) But Petitioner provided
`
`several reasons to combine Bergen and Reimer so that Bergen’s functionality could
`
`be provided during a video. (Pet., 41-42.) Moreover, Petitioner explained that a
`
`POSITA would have understood that Bergen’s current frame must be identified for
`
`Bergen’s system to work. (Id., 42-43.) To the extent not already disclosed by or
`
`inherent in Bergen, Petitioner explained that a POSITA would have understood
`
`that the necessary step of identifying the request location could have been
`
`performed as disclosed in Reimer. (Pet. 43-44.)
`
`2.
`
`The Combination Does Not Change Bergen’s
`Principle of Operation.
`
`PO argues the combination would require a change in Bergen’s “principle of
`
`operation” because Reimer “allows users to select scenes from already playing
`
`movies.” (POR, 36-38.) But Bergen also allows user requests during a video.
`
`(Ex. 1028, 20:49-54; POR, 41.) While PO characterizes Bergen’s system as
`
`“reactionary,” PO never explains why Bergen’s search could not be performed—
`
`and results displayed—in response to a query during a video as Reimer discloses.
`
`Petitioner’s argument does not depend on the physical combinability of the
`
`systems. (Ex. 1102 ¶¶51, 54, 58); Allied Erecting & Dismantling Co., 825 F.3d at
`
`1381.
`
`-17-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`B.
`
`Bergen Discloses the Retrieving Limitations.
`
`PO argues Bergen does not disclose the “retrieving” limitations because
`
`Bergen’s searching and output are more “comprehensive” than the claims require.
`
`(POR, 38-39.) The Board previously rejected this argument because it is not
`
`commensurate with the claims’ scope. (Paper 13 at 19; see also Ex. 1102 ¶66.)
`
`Although PO now cites expert testimony, that testimony does not support PO’s
`
`argument. (Ex. 2021 ¶140.)
`
`C. Bergen Discloses the Displaying Limitations.
`
`Neither PO nor Reader disputes that Bergen discloses providing annotations
`
`or scene descriptions, or that these constitute “information.” (Pet., 48.) Thus,
`
`Bergen discloses this limitation.
`
`PO argues Bergen’s “storyboard” images and “indicator of quality match”
`
`are not “information” because “information” has a more limited meaning in the
`
`’282 patent. (POR, 39-40.) That argument was contradicted by Reader, who
`
`admitted that “images representing other scenes in the video” are “information.”
`
`(Ex. 1101, 128:19-129:1, 119:20-122:7.)
`
` Reader further admitted
`
`that
`
`“information” is not limited to any particular type, format, or content. (Id.,
`
`117:15-118:17; see id. (“any information related to the frame number satisfies this
`
`limitation”), 126:15-129:11; Ex. 1102 ¶¶67-72.)
`
`-18-
`
`
`
`IPR2018-01498
`Amazon.com v. CustomPlay
`
`PO’s argument also contradicts the ’282 patent, which confirms that images
`
`are “information.” (Ex. 1001, 9:8-12, claim 11; id., 9:12-15 (“information” need
`
`not “provide any specific information”); Ex. 1102 ¶70.) Thus, PO’s attempt to
`
`exclude certain “informati