Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 1 of 27 PageID #: 2226
`
`
`IN THE UNITED STATES DISTRICT COURT
`FOR THE DISTRICT OF DELAWARE
`
`
`
`C.A. No. 23-1236 (GBW)
`
`
`
`
`NOKIA TECHNOLOGIES OY,
`
`Plaintiff,
`
`
`
`v.
`
`
`AMAZON.COM, INC., AND
`AMAZON.COM SERVICES LLC., AND
`TWITCH INTERACTIVE, INC.
`
`Defendants.
`
`
`
`
`
`PLAINTIFF’S OPPOSITION TO DEFENDANTS’ MOTION AND OPENING BRIEF IN
`SUPPORT OF MOTION TO DISMISS PURSUANT TO FED. R. CIV. P. 12(B)(6)
`
`
`Dated: March 15, 2024
`
`Warren H. Lipschitz (admitted pro hac vice)
`Alexandra F. Easley (admitted pro hac vice)
`MCKOOL SMITH, P.C.
`300 Crescent Ct. Ste. 1200
`Dallas, TX 75224
`Tel.: (214) 978-4000
`wlipschitz@mckoolsmith.com
`aeasley@mckoolsmith.com
`
`R. Mitch Verboncoeur (admitted pro hac
`vice)
`MCKOOL SMITH, P.C.
`303 Colorado St Suite 2100
`Austin, TX 78701
`Tel.: (512) 692-8700
`mverboncoeur@mckoolsmith.com
`
`Josh Newcomer (admitted pro hac vice)
`MCKOOL SMITH, P.C.
`600 Travis St., Suite 7000
`Houston, Texas 77002
`Tel.: (713) 485-7300
`jnewcomer@mckoolsmith.com
`
`Kevin Burgess (admitted pro hac vice)
`MCKOOL SMITH, P.C.
`
`Brian E. Farnan (Bar No. 4089)
`Michael J. Farnan ( Bar No. 5165)
`FARNAN LLP
`919 N. Market St. 12th Floor
`Wilmington DE 19801
`Tel.: (302) 777-0300
`bfarnan@farnanlaw.com
`mfarnan@farnanlaw.com
`
`Theodore Stevenson, III (admitted pro hac
`vice)
`ALSTON & BIRD LLP
`2200 Ross Ave. #2300
`Dallas, TX 75201
`Tel.: (214) 922-3400
`ted.stevenson@alston.com
`
`John D. Haynes (admitted pro hac vice)
`Nicholas T. Tsui (admitted pro hac vice)
`Shawn Gannon (admitted pro hac vice)
`ALSTON & BIRD LLP
`1201 West Peachtree Street
`Atlanta, GA 30309
`Tel.: (404) 881-7000
`john.haynes@alston.com
`nick.tsui@alston.com
`shawn.gannon@alston.com
`
`
`
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 2 of 27 PageID #: 2227
`
`104 East Houston St., Suite 300
`Marshall, Texas 75670
`Tel.: (903) 923-9000
`kburgess@mckoolsmith.com
`
`Attorneys for Plaintiff
`
`ii
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 3 of 27 PageID #: 2228
`
`
`I. 
`
`II. 
`
`III. 
`
`IV. 
`
`V. 
`
`
`
`TABLE OF CONTENTS
`
`
`STATEMENT OF THE NATURE AND STAGE OF THE
`PROCEEDINGS ................................................................................................................. 1 
`
`SUMMARY OF ARGUMENT .......................................................................................... 1 
`
`STATEMENT OF FACTS ................................................................................................. 2 
`A. 
`The ’469, ’599, and ’273 Patents ............................................................................ 3 
`B. 
`The ’818 Patent ....................................................................................................... 4 
`C. 
`The ’005 and ’764 Patents ...................................................................................... 5 
`
`B. 
`
`ARGUMENT ...................................................................................................................... 6 
`A. 
`The ’469, ’599, and ’273 Patents are patent-eligible .............................................. 6 
`1. 
`The ’469, ’599, and ’273 Patents are not abstract. ...................................... 6 
`2. 
`The ’469, ’599, and ’273 Patents claim inventive concepts. ...................... 8 
`The ’818 Patent is patent-eligible ........................................................................... 9 
`1. 
`The ’818 Patent is not abstract. ................................................................... 9 
`2. 
`The ’818 Patent claims inventive concepts. .............................................. 13 
`3. 
`Claim 1 is not representative. .................................................................... 15 
`The ’005 and ’764 Patents are patent-eligible ...................................................... 15 
`1. 
`The ’005 and ’764 Patents are not abstract. .............................................. 15 
`2. 
`The ’005 and ’764 Patents claim inventive concepts. ............................... 18 
`3. 
`Dependent Claims are Patentable for Additional Reasons. ...................... 20 
`
`C. 
`
`CONCLUSION ................................................................................................................. 20 
`
`
`
`i
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 4 of 27 PageID #: 2229
`
`TABLE OF AUTHORITIES
`
`
`
`Page(s)
`
`Cases
`
`Aatrix Software, Inc. v. Green Shades Software, Inc.,
`882 F.3d 1121 (Fed. Cir. 2019)................................................................................2, 14, 18, 19
`
`ADASA Inc. v. Avery Dennison Corp.,
`55 F.4th 900 (Fed. Cir. 2022) ......................................................................................12, 16, 18
`
`BASCOM Glob. Internet Servs., Inc. v. AT&T Mobility LLC,
`827 F.3d 1341 (Fed. Cir. 2016)....................................................................................13, 18, 19
`
`Berkheimer v. HP, Inc.,
`881 F.3d 1360 (Fed. Cir. 2018)..............................................................................................1, 2
`
`Cellspin Soft, Inc. v. Fitbit, Inc.,
`927 F.3d 1306 (Fed. Cir. 2019)................................................................................................19
`
`Cleveland Clinic Found. v. True Health Diagnostics LLC,
`859 F.3d 1352 (Fed. Cir. 2017)................................................................................................15
`
`Cooperative Ent’mt Inc. v. Kollective Tech., Inc.,
`50 F.4th 127 (Fed. Cir. 2022) ........................................................................................1, 13, 14
`
`Cosmokey Sols. GMBH & Co. KG v. Duo Sec. LLC,
`15 F.4th 1091 (Fed. Cir. 2021) ..........................................................................................13, 18
`
`Enfish LLC v. Microsoft Corp.,
`822 F.3d 1327 (Fed. Cir. 2016)..............................................................................10, 11, 12, 16
`
`Finjan, Inc. v. Blue Coat Sys.,
`879 F.3d 1299 (Fed. Cir. 2018)..........................................................................................10, 16
`
`Koninklijke KPN N.V. v. Gemalto M2M GmbH,
`942 F.3d 1143 (Fed. Cir. 2019)............................................................................................8, 17
`
`McRo Inc. v. Bandai Namco Games Am. Inc.,
`837 F.3d 1299 (Fed. Cir. 2016)................................................................................................17
`
`OnDemand LLC v. Spotify Tech., S.A.,
`
`484 F. Supp. 3d 188, 207 (D. Del. 2020) .................................................................................20
`
`Packet Intelligence LLC v. NetScout Sys.,
`965 F.3d 1299 (Fed. Cir. 2020)..........................................................................................12, 16
`
`ii
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 5 of 27 PageID #: 2230
`
`Peloton Interactive, Inc. v. Echelon Fitness, LLC,
`2020 U.S. Dist. LEXIS 118945, 2020 WL 3640064 (D. Del. July 6, 2020) ...........................14
`
`Rapid Litig. Mgmt. Ltd. v. CellzDirect, Inc.,
`827 F.3d 1042 (Fed. Cir. 2016)................................................................................................19
`
`RecogniCorp, LLC v. Nintendo Co.,
`855 F.3d 1322 (Fed. Cir. 2017)..................................................................................................7
`
`Throughtek Co., Ltd. v. Reolink Innovation Inc. et al.,
`2024 U.S. Dist. LEXIS 20139 (D. Del. Feb. 5, 2024) .................................................12, 14, 17
`
`TLI Commc’ns LLC v. AV Auto., L.L.C. (In re TLI Commc’ns LLC Patent Litig.),
`823 F.3d 607 (Fed. Cir. 2016)........................................................................................6, 11, 17
`
`Statutes
`
`35 U.S.C. § 101 ...................................................................................................................... passim
`
`Other Authorities
`
`Rule 12 ...................................................................................................................................1, 6, 13
`
`
`
`iii
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 6 of 27 PageID #: 2231
`
`I.
`
`STATEMENT OF THE NATURE AND STAGE OF THE PROCEEDINGS
`
`Nokia Technologies Oy (“Nokia”) filed its 591-page Original Complaint on October 31,
`
`2023, alleging, among other things, infringement of 15 patents by Amazon. Nokia’s complaint
`
`details the groundbreaking novelty of Nokia’s patented technologies and Amazon’s corresponding
`
`rampant infringement and flagrant misconduct. In response, Amazon filed its partial motion to
`
`dismiss alleging that six of the 15 asserted patents are invalid under 35 U.S.C. § 101 (D.I. 18).
`
`Amazon’s motion should be denied. Amazon attacks strawmen revisions of Nokia’s
`
`patents
`
`rather
`
`than
`
`the actual, claimed
`
`innovations. Amazon’s oversimplifications,
`
`mischaracterizations, and cursory treatment of the six challenged patents falls far short of
`
`Amazon’s burden of showing by clear and convincing evidence that the challenged claims are
`
`patent ineligible, particularly given the Federal Circuit’s repeated recognition that “[d]etermining
`
`whether the [invention] is well-understood, routine, or conventional is a question of fact that
`
`cannot be resolved at the Rule 12(b)(6) stage.” See, e.g., Cooperative Ent’mt Inc. v. Kollective
`
`Tech., Inc., 50 F.4th 127, 133 (Fed. Cir. 2022).
`
`II.
`
`SUMMARY OF ARGUMENT
`
`1.
`
`Patent eligibility is a question of law that depends on underlying issues of fact. See
`
`Berkheimer v. HP, Inc., 881 F.3d 1360, 1365 (Fed. Cir. 2018). “[P]atent eligibility may be resolved
`
`at the Rule 12 stage only if there are no plausible factual disputes after drawing all reasonable
`
`inferences from the intrinsic and Rule 12 record in favor of the non-movant.” Id.
`
`2.
`
`The challenged patents are not directed to abstract ideas. Instead, the challenged
`
`patents are directed to non-abstract technological advancements that are analogous to subject
`
`matter that the Federal Circuit has routinely found to be non-abstract and patent eligible. The Court
`
`may deny Amazon’s motion on this basis alone.
`
`1
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 7 of 27 PageID #: 2232
`
`3.
`
`The challenged patents are also patent eligible because each of the challenged
`
`patents contain an “inventive concept” as the claims are directed to specific technical solutions
`
`that were not well-understood, routine, or conventional at the time. Such well-pled allegations that
`
`claims are directed to an inventive concept are more than sufficient to deny Amazon’s motion.
`
`Aatrix Software, Inc. v. Green Shades Software, Inc., 882 F.3d 1121, 1126-27 (Fed. Cir. 2019).
`
`4.
`
`Amazon treats claims of the challenged patents as representative when they are not.
`
`This is an additional reason to deny Amazon’s motion. See Berkheimer, 881 F.3d at 1365.
`
`III.
`
`STATEMENT OF FACTS
`
`Each of the patents challenged by Amazon relates to specific technological solutions to
`
`problems arising in the field of video coding. The patents do not cover the concept of video coding,
`
`generically, as Amazon alleges. Instead, the challenged patents claim novel and improved ways of
`
`encoding or decoding video. The inventions are unique to video coding, without pencil and paper
`
`analogue, not capable of being accomplished in the human mind, and not directed towards any
`
`fundamental truth or economic practice. These improvements include: sub-pixel interpolation
`
`techniques, requiring fewer processing steps while enhancing accuracy; entirely new parameter
`
`set structures, aligning with the expected updating frequency of parameters in a bitstream, reducing
`
`data to be transmitted and allowing greater flexibility in parameter signaling; and reference picture
`
`tracking, enabling differentiation between reference and non-reference pictures in error situations.
`
`Digital video sequences, like movies streaming over the Internet, are made of sequences of
`
`pictures (called frames). The illusion of motion in a video is created by rapidly displaying the
`
`frames in sequence (’469 Patent, 1:11-17). Frames are formed by image pixels, which are the
`
`smallest addressable unit of a digital video frame and can have various degrees of brightness and
`
`color (id., 1:32-36). To facilitate compression and transmission of the pictures, each frame is
`
`subdivided into sections known as slices, which are themselves composed of blocks—regions of
`
`2
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 8 of 27 PageID #: 2233
`
`pixels which are encoded and decoded together (id., 1:61-2:5; D.I. 1, ¶ 64; ’818 Patent, 1:51-55).
`
`Because of the fast frame rate, images in consecutive frames are often similar and contain
`
`redundant information. For example, a block depicting a portion of background scenery in one
`
`frame may appear in the next frame at a slightly different location as the camera pans. This is
`
`called “temporal redundancy,” i.e., referring to portions of the image that do not change
`
`substantially from one picture to the next (’469 Patent, 1:11-31). Video compression methods take
`
`advantage of temporal redundancy by coding information that allows a video playback device
`
`(decoder) to predict the content of a block of pixels in a current frame based on a previous frame,
`
`the latter of which is called a “reference frame” or “reference picture” (D.I. 1, ¶¶ 133, 144; ’005
`
`Patent, 1:14-19). Prediction based on reference frames significantly reduces the need to transmit,
`
`store, or receive redundant information, thus saving bandwidth from pixel values that will not need
`
`to be re-transmitted (see ’469 Patent, 1:11-31, 2:37-56, 3:37-59).
`
`A.
`
`The ’469, ’599, and ’273 Patents
`
`The ’469, ’599, and ’273 Patents are in the same patent family. Each is directed to patent-
`
`eligible subject matter. In the interest of brevity, Nokia responds with respect to the ’469 Patent as
`
`representative of each of its family members.1 The ’469, ’599, and ’273 Patents are directed to an
`
`improved technique for computing interpolated subpixel values for use in motion compensated
`
`video coding (see ’469 Patent, 1:4-7). For example, in motion compensated video coding, a motion
`
`vector indicates how a block of pixels from a reference frame should move when predicting the
`
`next block of pixels (see id., 3:37-4:15). By transmitting a motion vector instead of an entire block
`
`of pixels representing the next frame, the amount of information needed to represent the sequence
`
`
`1 Nokia does not concede the claims of the ’469 Patent are substantially the same as the claims of
`the ’599 and ’273 Patents. However, all three patents are patent-eligible for similar reasons.
`
`3
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 9 of 27 PageID #: 2234
`
`is reduced and bandwidth is saved (see id.). However, blocks do not always move by a whole
`
`number of pixels, so motion vectors are often represented with subpixel accuracy, allowing them
`
`to point “in-between” pixels of the original picture (see id., 6:20-48). Because pixels are the
`
`smallest addressable unit of digital video, the values of these subpixels are determined by
`
`interpolating between the values of pixels in the reference block, a computationally intensive task
`
`(see id., 6:20-48, 7:55-62).
`
`Prior to the ’469 Patent, two test models—TML5 and TML6—were proposed for
`
`interpolating between pixels to arrive at sub-pixel values, but both imposed increased complexity
`
`and cost on devices. For example, both proposals increased the number of arithmetic operations or
`
`required greater hardware complexity for achieving those operations with requisite precision (see
`
`id., 8:43-53, 11:14-32, 13:20-29). The ’469 patent resolves these limitations by providing “a
`
`method for sub-pixel value interpolation capable of providing satisfactory performance in both the
`
`encoder and decoder” in a way which neither TML5 nor TML6 could (see id., 13:30-39).
`
`B.
`
`The ’818 Patent
`
`Video encoders and decoders rely on different parameters that specify the encoding and
`
`decoding of video data. Transmission, storage, and reception of these parameters affects
`
`bandwidth, memory, and computational demands. In one approach, prior to the ’818 Patent,
`
`parameters relevant to a given “slice” of video were transmitted in a slice header, while sequence
`
`parameters and picture parameters were transmitted in a single parameter set (see, e.g., ’818 Patent,
`
`2:48-56). A significant problem with this prior art single parameter set concept was that in order
`
`to signal a single different parameter value, the entire parameter set had to be transmitted again
`
`even though all of the other parameter values were the same (D.I. 1, ¶ 64). Transmitting and
`
`receiving multiple substantially identical parameter sets was inefficient and caused latency and
`
`reliability issues (id.; ’818 Patent, 3:17-28). The’818 Patent claims a specific technical
`
`4
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 10 of 27 PageID #: 2235
`
`improvement to those prior video coding methods by providing a novel data structure of distinct
`
`sequence and picture parameter sets, where “parameters which are not allowed to change in a
`
`coded video sequence are included in the sequence parameter set,” parameters that are allowed to
`
`change in every picture but likely changed only every few pictures were included in a separate
`
`picture parameter set, and “[p]arameters whose value may change in every slice or whose value is
`
`likely to change in every picture, are included in the slice header” (see, e.g., ’818 Patent, 4:15-29).
`
`The ’818 Patent claims a technical solution to problems specific to coded sequences of
`
`pictures in a bitstream and improves video coder functionality. The ’818 Patent claims use of an
`
`improved data structure for transmission (and reception) of sequence and picture parameters by
`
`“split[ting] the parameter set structure [in]to multiple parameter set structures according to the
`
`persistency and target of the parameters,” with at least one picture parameter value remaining
`
`unchanged in all slice headers of a picture (D.I. 1, ¶ 66; ’818 Patent, 4:15-19). The invention results
`
`in more efficient transmission of parameter information and increased compression efficiency
`
`within an encoder and/or a decoder of streaming video (D.I. 1, ¶¶ 66-67; ’818 Patent, 6:50-59).
`
`C.
`
`The ’005 and ’764 Patents
`
`As discussed above, video coding techniques reduce the transmission of redundant
`
`information by predicting pictures from “reference pictures” that contain similar image content
`
`(see, e.g., ’005 Patent, 1:14-19). Any loss or error in a reference picture will propagate into every
`
`other picture that relies on that reference picture, which degrades the quality of the entire video
`
`(D.I. 1, ¶¶ 135, 146). Conventional video coding technology did not include information in the
`
`bitstream to distinguish reference pictures (used to predict other pictures) from non-reference
`
`pictures (not used to predict other pictures). The result was that when any picture was lost or
`
`corrupted, the system was required to freeze and wait for retransmission or expend resources
`
`attempting error concealment (D.I. 1, ¶¶ 136-137). There was no ability to recognize that the lost
`
`5
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 11 of 27 PageID #: 2236
`
`or corrupted picture was a non-reference picture, whose absence could be safely ignored because
`
`no other pictures were being predicted from it (id.).
`
`The ’005 and ’764 Patents claim technical solutions to this problem, which arises only in
`
`the context of coding sequences of pictures in a bitstream. Specifically, the claims are directed to
`
`a novel technique of separately tracking reference pictures in a bitstream, which enables the
`
`differentiation between a loss of a reference picture from a loss of a non-reference picture (D.I. 1,
`
`¶¶ 137, 148; ’005 Patent, 4:3-12, cls. 1, 5; ’764 Patent, cl. 1).2
`
`IV. ARGUMENT
`
`A.
`
`The ’469, ’599, and ’273 Patents are patent-eligible
`
`1. The ’469, ’599, and ’273 Patents are not abstract.
`
`As set forth in the Complaint, the ’469 Patent “is directed to novel and unconventional
`
`improvements to motion-compensated prediction in the field of digital video coding,” and
`
`specifically “provides improvements over prior motion compensated prediction and video
`
`compression techniques that result in substantial benefits to motion prediction, video compression,
`
`video quality, and video playback” (D.I. 1, ¶ 72). Amazon incorrectly alleges that the claims of
`
`the ’469 Patent “are directed to nothing more than the idea of calculating sub-pixel values using
`
`weighted sums and averages of other pixel and sub-pixel values,” and that these claims are directed
`
`to “in other words, the idea of encoding and decoding video using basic math” (D.I. 18., 11). This
`
`is a reductive mischaracterization and violates the tenant that a court should not oversimplify the
`
`claimed invention, particularly when considering a Section 101 challenge at the Rule 12 stage. TLI
`
`Commc’ns LLC v. AV Auto., L.L.C. (In re TLI Commc’ns LLC Patent Litig.), 823 F.3d 607, 611
`
`
`2 The ’005 and ’764 Patents share a specification but the independent claims are meaningfully
`distinct with respect to subject matter eligibility, as explained herein.
`
`6
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 12 of 27 PageID #: 2237
`
`(Fed. Cir. 2016) (citing Alice, 134 S. Ct. at 2354).
`
`The challenged claims relate to a specific improvement to sub-pixel interpolation
`
`techniques in video coding rather than a paint-by-numbers implementation of image
`
`encoding/decoding or generic interpolation techniques. Cf. RecogniCorp, LLC v. Nintendo Co.,
`
`855 F.3d 1322, 1326 (Fed. Cir. 2017). The claimed improved sub-pixel interpolation technique
`
`offers significant benefits over prior art (such as TML5 and TML6): It requires fewer processor
`
`cycles, reduces the number of transistors needed to implement it in hardware, and enhances
`
`accuracy by avoiding reliance on truncated intermediate values when interpolating sub-pixels (see
`
`’469 Patent, 37:33-38:51). Table 1 of the ’469 Patent demonstrates the measurable improvements
`
`of the claimed technique. The invention is computationally more efficient because, compared to
`
`the preceding TML5 and TML6 methodologies, the invention reduces the number of linear and 6-
`
`tap filtering operations necessary to compute the value of a sub-pixel at every sub-pixel location
`
`(see id., 41:64-67). It also requires less memory (see id., 40:8-38). This reduction in both
`
`computational and memory complexity directly reduces the cost and size of hardware used to
`
`practice the claims and leads to a tangible improvement in the functioning of the hardware (see
`
`id., 38:5-12, 38:21-24).
`
`Amazon relies on RecogniCorp, taking the position that “[t]he idea of encoding and
`
`decoding video is abstract as a matter of law” (D.I. 18, 11). Putting aside Amazon’s gross
`
`oversimplification of the claims, this overextends RecogniCorp, where the Federal Circuit found
`
`that a broad claim for creating a composite facial image based on a facial code “derived by
`
`performing at least one multiplication operation” was directed to the abstract idea of “encoding
`
`and decoding image data.” See RecogniCorp at 1326. This case does not, however, support the
`
`broad proposition that any claim relating in some way to video encoding or decoding is abstract.
`
`7
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 13 of 27 PageID #: 2238
`
`A later opinion by the Federal Circuit recognizes that the claims in RecogniCorp “did not
`
`adequately capture the inventors’ asserted technical contribution, because the claims recited no
`
`more than ‘standard encoding and decoding, an abstract concept long utilized to transmit
`
`information.’” Koninklijke KPN N.V. v. Gemalto M2M GmbH, 942 F.3d 1143, 1153 (Fed. Cir.
`
`2019). In contrast, the claims of the ’469 Patent recite specific solutions to technical problems with
`
`sub-pixel interpolation for use in motion compensation in video encoding and decoding. See
`
`Koninklijke KPN N.V. at 1153 (determining challenged claims “patent-eligible because they are
`
`directed to a non-abstract improvement in an existing technological process”).
`
`2. The ’469, ’599, and ’273 Patents claim inventive concepts.
`
`While the Court does not need to reach step two, Amazon’s step two analysis also
`
`oversimplifies the claims to avoid addressing their substance. Amazon contends that “[e]ach
`
`element of the claims simply recites an ‘interpolating’ step, without reciting any new hardware or
`
`special programming” (D.I. 18, 13). Not so. Prior to the ’469 patent, existing solutions for
`
`computing sub-pixel values, such as TML5, needed significant computational resources because
`
`sub-pixel values were heavily inter-dependent, meaning that calculating one sub-pixel’s value
`
`might necessitate the calculation of many other sub-pixels (see ’469 Patent, 41:4-10). Others, such
`
`as TML6, imposed significant memory burdens by requiring high-precision intermediate values
`
`be stored as precomputed sub-pixels before a final sub-pixel value could be calculated (see id.,
`
`37:36-38:24, 40:27-38). The ’469 Patent, by contrast, claims an inventive approach to sub-pixel
`
`interpolation that balances reduction of computational complexity with accuracy by, for example,
`
`calculating all ¼ resolution sub-pixel values based on linear interpolation values from original
`
`pixels or ½ pixel values (see id., 37:39-41, Table 1 (reduced computational complexity)).
`
`Far from claiming the use of “interpolation” without substantiation, as Amazon alleges, the
`
`claims describe how to arrive at ¼ sub-pixel values without the detriments of competing
`
`8
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 14 of 27 PageID #: 2239
`
`techniques. For example, claim 1 recites:
`
`c) when a value for a sub-pixel situated at a ½N unit horizontal and
`½N unit vertical location is required, interpolating such a value by
`taking a weighted average of the value of a first sub-pixel or pixel
`situated at a ½N−m unit horizontal and ½N−n unit vertical location and
`the value of a second sub-pixel or pixel located at a ½N−p unit
`horizontal and ½N−q unit vertical location, variables m, n, p and q
`taking integer values in the range 1 to N such that the first and
`second sub-pixels or pixels are located diagonally with respect to
`the sub-pixel at ½N unit horizontal and ½N vertical location.
`
`(Id., Cl. 1). The claim articulates a specific algorithm by which to select the sub-pixels or
`
`pixels from which to interpolate other sub-pixel values. This is substantially more than generic
`
`interpolation and is a specific improvement to pre-existing interpolation techniques. The claim
`
`allows a reduction in the precision of the arithmetic required, making the computation substantially
`
`easier. This increases the speed with which the operations can be performed while also lowering
`
`the cost of hardware required (see id., 38:15-24). Further, this technique enables flexibility in the
`
`computation of sub-pixel values because they can be determined by interpolation in either the
`
`horizontal or vertical directions, depending on which sub-pixel value is required (see id., 38:34-
`
`51, cl. 1 at step b)).
`
`Taken as an ordered combination in their entirety, the steps of each claim in the ’469 Patent
`
`are inventive. While TML5 and TML6 arrive at sub-pixel values for motion compensation, each
`
`step of the claims at issue describes a specific association between a sub-pixel and neighboring
`
`pixel and subpixel values in a way that improves efficiency while enhancing accuracy. Amazon’s
`
`overly broad summarization of each such step as “interpolation” is irrelevant in the face of the
`
`actual claim language and does not render the claims non-inventive.
`
`B.
`
`The ’818 Patent is patent-eligible
`
`1. The ’818 Patent is not abstract.
`
`The claims of the ’818 Patent are directed to specific technical improvements to video
`
`9
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 15 of 27 PageID #: 2240
`
`encoding and decoding: namely a new method and device for video coding that utilizes sequence
`
`parameter sets (for parameters that are constant throughout a sequence), picture parameter sets (for
`
`parameters that likely change every few pictures but can change for each picture) and slice headers
`
`(for parameters likely to change from picture to picture and may change from slice to slice within
`
`a given picture) (D.I. 1, ¶¶ 66-67; ’818 Patent, 4:15-29, 6:50-59, 7:64-9:6). This novel concept—
`
`which is an improvement over a single parameter set—results in increased efficiency and
`
`compression (see, e.g., ’818 Patent, 4:15-29, 6:50-59, 7:64-9:6). The claimed separation of
`
`sequence parameter sets, picture parameter sets, and inclusion of specific parameters in slice
`
`headers was a specific improvement in computer (coding) functionality over conventional
`
`technology that had only one parameter set and required repeated signaling of many unchanged
`
`parameters any time even a single parameter value changed (see, e.g., ’818 Patent, 6:50-7:5, 7:39-
`
`45, 3:56-58; D.I. 1, ¶¶ 66-67). Additionally, the claimed “at least one picture parameter value in a
`
`slice header” further results in reduced video file sizes because including every parameter “whose
`
`value is likely to change in every picture” in the picture parameter set would require a unique
`
`picture parameter set for every individual picture (see, e.g., ’818 Patent, 4:15-26). Including at
`
`least one picture parameter value in the slice header, as in the claims and specification, eliminates
`
`the need to repeat the picture parameter set for each picture, which in turn reduces the signaling of
`
`unchanged parameters (see, e.g., ’818 Patent, 4:21-24, 7:39-42).
`
`Thus, the ’818 Patent claims multiple novel improvements, each precisely the type of
`
`“specific implementation of a solution to a problem in the software arts” that the Federal Circuit
`
`has held are not abstract. Enfish LLC v. Microsoft Corp., 822 F.3d 1327, 1336 (Fed. Cir. 2016) (“a
`
`specific type of data structure designed to improve the way a computer stores and retrieves data in
`
`memory” is not abstract); Finjan, Inc. v. Blue Coat Sys., 879 F.3d 1299, 1305 (Fed. Cir. 2018)
`
`10
`
`

`

`Case 1:23-cv-01236-GBW Document 22 Filed 03/15/24 Page 16 of 27 PageID #: 2241
`
`(improvements to computer functionality “to do things it could not do before” are not abstract).
`
`The ’818 Patent eliminated the need to transmit an entirely new parameter set for a change to a
`
`single picture parameter value using novel signaling techniques, achieving improved compression
`
`efficiency over conventional technology (see, e.g., ’818 Patent, 6:50-59, 7:39-42, 4:15-26; D.I. 1,
`
`¶¶ 69-70). Enfish, 822 F.3d at 1337 (“our conclusion … is bolstered by the specification’s
`
`teachings that the claimed invention achieves other benefits over conventional [technology]”).
`
`Rather than addressing the technical improvements offered by the ’818 Patent, Amazon
`
`instead asserts that these specific technical solutions—the claimed signaling of video coding
`
`parameters into a sequence parameter set, picture parameter set, and slice header based on the
`
`expected frequency of change in the parameter—could somehow be performed by a newborn
`
`looking at pictures for similarities (D.I. 18, 18). Amazon fails to describe any serious parallels
`
`between “an innate human characteristic” (id.) and the signaling of video coding parameters using
`
`different techniques based on the frequency of change of the parameter as claimed in the ’818
`
`Patent. This is just the sort of extreme oversimplification the Federal Circuit has repeatedly
`
`counseled against. TLI Commc’ns, 823 F.3d at 611. Amazon’s other overly simplistic analogy—
`
`classifying photos according to date—has nothing to do with parameter values used in encoding
`
`and decoding pictures for bit savings, does not consider how often different parameter values are
`
`expected to change, and could not achieve any of the data compression benefits of the ’818 Patent.
`
`Amazon’s dramatic oversimplifications highlight that there are no pen and paper analogues to the
`
`claimed invention, much less