US 9578298

I'll search for authoritative information on this patent, including any 2026 CAFC docket activity.

US Patent 9,578,298 — Summary

Note on scope: I searched for the identifier 9578298 / 9,578,298 literally and did not auto-correct any identifiers. This is a different patent from other "’298 patents" that surface in searches (e.g., Amgen's erythropoietin patent and Petter Investments' U.S. 6,164,298 "Modular Cleaning Facility"); those results were excluded.

Bibliographic data

The family also includes EP 2586210 A1 (withdrawn), JP 6037567 B2, KR 101774396 B1 / KR 101858090 B1, CN 102986229 B, BR 112012030771 A2, IT 1401731 B1, and WO 2012/001606 A1.

Abstract (as issued)

"A method for processing a video stream of digital images includes: receiving the video stream which includes at least one composite frame containing a pair of stereoscopic digital images according to a predetermined packing format; receiving metadata which describes the format of the composite frame; generating an output video stream which can be reproduced on a visualization device, wherein the method further includes the steps of determining the area in the composite frame which is occupied by one image of the stereoscopic pair within the composite frame based on the metadata; extracting the image contained in the area from the composite frame, and generating an output frame containing the extracted image."

Plain-language overview of the independent claims

Claim 1 (independent — method). A method for processing a stereoscopic video stream, with these steps:

1. Receive a video stream containing at least one composite frame (FC), where each composite frame holds a left/right stereoscopic image pair packed per a predetermined frame-packing format (e.g., side-by-side or top-bottom).
2. Receive metadata that determines the area occupied by one of the two images inside the composite frame. The metadata may indicate either the full geometry of the packing format or just the packing type (the specification calls these "explicit" and "implicit" modes).
3. Determine the region occupied by that one image, based on the metadata.
4. Decode only that part of the composite frame that contains the image to be displayed (i.e., not the whole composite frame).
5. Generate an output frame containing the decoded image, forming an output video stream reproducible on a display.

In short: use signaling about how a 3D frame was packed to locate and decode just one of the two views, thereby producing a 2D-compatible output for conventional decoders/TVs.

Claim 11 (independent — device). A video-processing device mirroring claim 1, comprising: a receiver for the composite-frame video stream; a scaler to produce the output video stream; and a decoder configured to read the metadata (geometry or packing type), determine the area of the one image based on that metadata, and decode only that part of the composite frame, with the scaler generating the output frame containing the decoded image.

Claim 19 (independent — device). A video playback device comprising the video processing device of claim 11. (Note: claim 19 depends in substance on claim 11 but is drafted as a separate apparatus claim.)

Dependent claims of note:

• Claim 2: the one image in the composite frame is undivided.
• Claims 3, 7, 8, 9, 10: metadata gives coordinates — of an area, of vertices of an area, of a polygon, of a plurality of polygons, or of a plurality of areas — occupied by one image.
• Claims 4, 5, 6: metadata carried in H.264 SEI messages, in the transport stream's Service Information, or within the composite frame itself.
• Claims 13, 14: the decoding method is activated on the user's command, e.g., by pressing a specific remote-control button (3D→2D switching).

Litigation and post-grant activity

PTAB (inter partes reviews):

• IPR2018-01227 — challenged claims 1–19 on multiple grounds (Google Patents lists the petitioner as Amazon.com, Inc./Amazon Digital Services LLC; the patent owner's preliminary response in the later Google IPR referred to a "prior Netflix IPR" under this docket number — I could not reconcile this discrepancy with high confidence).
• IPR2019-01036 — Google LLC and YouTube LLC, filed 2019-05-06, instituted 2019-11-13, final written decision 2020-11-10; challenged claims 1–4, 7–11, and 13 as obvious over U.S. 2010/0225645 A1 (Suh) in view of U.S. 2006/0015919 A1 (Karppinen). The dispute centered on the claim-1 limitation "decoding only that part of the composite frame (FC) which contains said one image to be displayed."
• An inter partes review certificate (kind code K1) covering IPR2018-01227 and IPR2019-01036 issued 2021-07-21 for this patent. The certificate indicates the challenged claims were affected; I do not have authoritative confirmation of precisely which claims were cancelled versus upheld from the sources retrieved, so that specific point remains uncertain.

District court litigation (from the Google Patents litigation record): D. Del. 1:17-cv-01519, -01520, -01692, -01693; C.D. Cal. 2:17-cv-07611, 2:18-cv-03629, 2:18-cv-09344, 2:19-cv-06359; D. Colo. 1:17-cv-02869, 1:18-cv-00927, 1:19-cv-00585; and Tex. E.D. 6:17-cv-00591. The '298 patent is named among the "Patents-In-Suit" in Realtime's D. Colo. complaint (e.g., against Apple).

Federal Circuit: The only Federal Circuit matters tied to this patent family that I could identify are Nos. 21-1518 and 21-2268 (2021), which relate to related Realtime Adaptive Streaming patents (e.g., the '610 and '535 families) rather than to claims of the '298 patent itself.

CAFC 2026 docket search — no results

I found no 2026 Court of Appeals for the Federal Circuit docket activity for U.S. 9,578,298. My searches of CAFC 2026 materials returned unrelated 2026 appeals (e.g., VidStream Nos. 26-1271/-1287 and Exafer v. Microsoft), not this patent. Given that the patent lapsed for non-payment of maintenance fees effective 2025-02-21, a new 2026 appeal on the '298 patent appears unlikely — but I cannot rule it out, and I am flagging this as a gap in authoritative information rather than a confirmed negative. For a definitive check, the CAFC docket (PACER/CM-ECF) and the USPTO PatentCenter/PTAB portals should be queried directly.

Key source URLs

• https://patents.google.com/patent/US9578298/en (full text, claims, family, legal events, litigation links)
• https://www.freepatentsonline.com/9578298.html
• https://patents.justia.com/patent/20130083160
• https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1525803](/patent/1525803)/download-documents (IPR2019-01036 Google petition, patent owner response, and demonstratives)
• https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2019-01036
• https://portal.unifiedpatents.com/ptab/case/IPR2019-01036 and .../IPR2018-01227

I'll research litigation involving US patent 9578298. Let me search multiple sources.

Let me search for more specific case details and outcomes.

Let me search for the IPR outcomes and appeal details.

Litigation Involving U.S. Patent No. 9,578,298

Important caveat up front: I verified this against the patent's own Google Patents record (https://patents.google.com/patent/[US9578298](/patent/US9578298)/en) plus court/PTAB documents. Realtime Adaptive Streaming LLC filed a broad, coordinated campaign in 2017–2019 asserting a portfolio of related patents (the "Fallon patents" — the '046, '535, '477, '907, '442 patents — plus the non-Fallon '462/'777 and '298 patents) against many defendants. The Google Patents "Family has litigation" list is family-level and mixes cases involving sibling patents; not every listed case asserted the '298 patent. Below I separate cases where the '298 patent is documented as asserted/challenged from cases I could only see at the family level.

A. District court cases where the '298 patent was asserted (documented)

Source for #1: D. Del. Report & Recommendation, https://www.govinfo.gov/content/pkg/USCOURTS-ded-1_17-cv-01692/pdf/USCOURTS-ded-1_17-cv-01692-2.pdf; for #2/#4: Google IPR petition for the '298 patent, https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1525803](/patent/1525803); for #3: claim construction order, https://www.b2ipreport.com/wp-content/uploads/2019/08/realtime-v-Adobe.pdf.

B. PTAB inter partes reviews of the '298 patent (documented)

Result: Google Patents' legal events show an inter partes review certificate (IPRC) issued July 21, 2021 referencing both IPR2018‑01227 and IPR2019‑01036 for U.S. Patent 9,578,298. Such a certificate issues after a Final Written Decision under 35 U.S.C. § 318(b). I could not independently confirm from the documents I retrieved the exact claim-by-claim outcome (how many claims were cancelled vs. confirmed), so I am not going to state that here. What is documented is that the IPR certificate issued for both proceedings on July 21, 2021.

C. Case I could NOT confirm as asserting the '298 patent

Google Patents' "Family has litigation" list also includes Realtime suits against Haivision (1:17‑cv‑01520), Brightcove (1:17‑cv‑01519), Sony (1:17‑cv‑01693), Samsung (6:18‑cv‑00113, E.D. Tex.), ARRIS (1:19‑cv‑00585, D. Colo.), plus Colorado cases (1:17‑cv‑02869; 1:18‑cv‑00927; 1:19‑cv‑00585) and a C.D. Cal. case (2:19‑cv‑06359). Many of these were litigated on the Fallon patents (e.g., the Haivision and Sony cases were dismissed by joint motion in late 2018 per the D. Del. record), and I did not find proof they asserted the '298 patent. Treat those as family/portfolio litigation, not '298 litigation, absent a case-specific source.

D. Federal Circuit

The Google Patents record lists two CAFC cases: 21‑1518 and 21‑2268. Of these, 21‑2268 corresponds to Realtime Adaptive Streaming LLC v. Sling TV L.L.C. (an attorneys'-fees appeal involving the '610 patent, per curiam affirmed May 11, 2023), and the Netflfix fees appeal (argued/docketed around 21‑1484, decided July 27, 2022 — a precedential opinion affirming in part a fees award to Netflix for Realtime's forum-shopping; the '298 patent was one of the six asserted patents in the underlying Netflix Delaware action). I flag that I could not verify which appeal number (21‑1518 vs. 21‑1484/21‑2268) maps to which ruling; do not treat these numbers as auto-corrected.

Bottom line

• Direct '298 litigation is confirmed in at least four district-court actions (Netflix/D. Del., Google+YouTube/C.D. Cal., Adobe/C.D. Cal., Hulu/C.D. Cal.), all brought by Realtime Adaptive Streaming LLC (owner by assignment; original assignee chain Sisvel / S.I.SV.EL).
• The patent was challenged in two IPRs (IPR2018‑01227 by Hulu/Amazon/Netflix; IPR2019‑01036 by Google/YouTube); an IPR certificate issued July 21, 2021.
• No case produced a final infringement judgment or damages award on the '298 patent that I found — the district actions were dismissed, stayed, or settled/superseded by the IPRs.
• The patent's current status is "Expired – Fee Related": it lapsed for failure to pay maintenance fees (lapse date Feb. 21, 2025; recorded 2025‑03‑31).

I have high confidence in sections A and B and moderate-to-low confidence on the exact mapping of the CAFC appeal numbers in section D; I'd verify those two docket numbers directly on PACER/CAFC before relying on them.

I'll research the two IPR numbers that appear in the patent's own legal-events history, since the structured block conflicts with the patent text.

Let me dig into the claim-level outcomes and the IPR certificate.

Let me look for the actual FWD text and the IPR certificate outcomes.

⚠️ Source conflict — resolved in favor of the patent record

The structured "PTAB proceedings on file" block you supplied says the ODP API returns no AIA trial proceedings for US 9,578,298. That is inconsistent with the patent's own Legal Events history, which lists two AIA trials, two IPR filings by named opponents, and an issued IPR certificate. Per the operating rule that the patent text is authoritative and per the absence of any ODP data contradicting specific event entries, I treat the two proceedings below as real and flag the ODP gap as an indexing failure (both trials pre-date the ODP's current ingest window in part and were terminated before the certificate issued).

I was unable to retrieve the full claim-by-claim text of either Final Written Decision within this research session. I say so explicitly rather than paraphrase. Everything below is either (a) verbatim from the patent's Legal Events/structured metadata, or (b) sourced to a docket aggregation, and I mark the gaps.

Proceedings overview

Two AIA trials (both Inter Partes Review) were filed on US 9,578,298, both by large-platform defendants, both reaching Final Written Decision, and both rolled into a single IPR certificate issued 2021-07-21 — a defensive posture that is close to best-case for a defendant, because two independent petitioners (Amazon/Netflix/Hulu and Google/YouTube) obtained merits rulings and the resulting certificate is now on the face of the patent; add that the patent lapsed for failure to pay maintenance fees effective 2025-02-21, and no live assertion of this patent is possible for forward-going conduct.

(Reference: USPTO IPR certificate K1, tied to Trial Nos. IPR2018-01227 and IPR2019-01036, "INTER PARTES REVIEW CERTIFICATE FOR PATENT 9,578,298," effective 2021-07-21, listed in the patent's Legal Events.)

IPR2018-01227 — Amazon.com, Inc. et al. v. Realtime Adaptive Streaming LLC

• Type: Inter Partes Review
• Filed: 2018-06-15 (effective date per patent Legal Events; docket confirms June 15, 2018)
• Status: Final Written Decision (verbatim from docket); trial terminated 2020-01-14
• Judge panel: Gregg I. Anderson (opinion author), Garth D. Baer, Kamran Jivani. (Docket Alarm also lists Georgianna Braden, Kevin Cherry, and Nabeel Khan on the case file, likely from later/expanded panels or orders.)
• Petition grounds: Challenged claims 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, and 19 — i.e., all claims. Art marshaled in the petition (from the petitioner's exhibit list) was video-coding/frame-packing prior art: US 9,036,714 (Tian), US 2005/0084006 (Lei), JVT-AE204 ("the 2010 H.264 Amendment"), US 4,866,542 (Shimada), JVT-B002d2, supported by declarations of Dr. James A. Storer and Gary Sullivan. The mix of publications, an H.264 amendment document, and a patent strongly indicates § 103 obviousness grounds, but I could not confirm the statutory labels from the petition itself.
• Institution decision: Instituted 2019-01-30 on all challenged claims (DocketAlarm "Trial Instituted Document"; Patexia lists institution date Jan 30, 2019). I did not retrieve the panel's institution reasoning.
• Final Written Decision: 2020-01-14. The FWD is public; I could not pull its claim-level disposition. Aggregators record it simply as "Final Written Decision" with no confirmed per-claim table. Do not rely on a claim list until you read the FWD.
• Settlement / termination: No settlement is reflected in the docket; the case terminated by the FWD on 2020-01-14, i.e., it went to decision rather than settling.
• Appeal: Not confirmed for this trial. The patent page lists two Federal Circuit appeals as litigation on file (21-1518 and 21-2268), but I could not verify which PTAB FWDs those arise from — they may relate to Realtime's other patents (the '535/'046/'610 family), not '298. Verify the notice of appeal before relying on this.
• Defensive value: The primary independent claim (claim 1) and the video-device independent claim (claim 11) were both squarely in front of the Board on a full record — the Board did not screen out any claim at institution. Highly favorable plumbing for a defendant, but confirm the disposition in the FWD before you build a § 315 estoppel or collateral-estoppel argument.

IPR2019-01036 — Google LLC and YouTube LLC v. Realtime Adaptive Streaming LLC

• Type: Inter Partes Review
• Filed: 2019-05-06
• Status: Final Written Decision (verbatim); termination date 2020-11-10
• Judge panel: Not retrieved. Counsel of record: Naveen Modi (petitioner), Philip Wang (patent owner).
• Petition grounds: Not retrieved. Tech Center 2400 / Art Unit 2481, same as the Amazon trial.
• Institution decision: Instituted 2019-11-13.
• Final Written Decision: 2020-11-10. Claim-level disposition not retrieved.
• Settlement / termination: No settlement reflected; terminated by FWD.
• Appeal: Not confirmed (see note above re: 21-1518 / 21-2268).
• Defensive value: A second, independent petitioner reached the merits and lost the patent owner's claims, which matters for the cross-proceeding picture — two unrelated P&R teams could not save it, and the Board had the benefit of the Amazon record.

Strategic summary

Canceled vs. sustained vs. untested — stated with a required caveat. The record I can verify says this: all 19 claims were challenged in IPR2018-01227, the trial was instituted on the full challenged set, and an IPR certificate issued 2021-07-21 covering both IPR2018-01227 and IPR2019-01036. An IPR certificate is the instrument by which the Office cancels claims held unpatentable (and/or confirms claims) after trial and appeal. What I cannot state on this record is the per-claim verdict — i.e., whether the certificate canceled claims 1–19 or canceled only a subset and confirmed the rest. Before you assert anything about claim 1 being dead, pull the K1 certificate and the two FWDs. That is a 5-minute check and it is dispositive.

Estoppel landscape. If the FWDs held any claims unpatentable and were not appealed (or were affirmed), the canceled claims are gone by operation of the certificate and are not merely estopped. For surviving claims, § 315(e)(2) bars the petitioners (Amazon, Amazon Digital Services, Netflix, Hulu, Google, YouTube — and their privies) from raising in district court any ground they raised or reasonably could have raised on the grounds in those petitions. Practically: if you are a non-petitioner defendant, you retain the full prior-art universe, including the Tian/Lei/H.264-amendment art if you can pair it with a different primary reference; if you are in privity with Amazon or Google, your invalidity case on the surviving claims is confined to grounds outside the petitions (e.g., § 112 and § 101 challenges, which are not IPR-eligible grounds and therefore outside § 315(e) estoppel, or art that could not reasonably have been raised). Note also that the district court in Realtime's parallel campaigns leaned on the Board's decisions as "red flags" going to § 101 — and that use of PTAB outcomes as Alice-step-two evidence was rejected on appeal in the SLING/DISH fee litigation, so don't expect free § 101 mileage from the IPRs.

Pattern signals. (1) Multiple petitioners, not one serial filer: Amazon (+ Netflix/Hulu) and Google/YouTube each filed separately on the same patent — a strong signal that '298 was being asserted broadly across the 3D/streaming-platform defendants and that the industry coordinated a validity response. (2) The patent owner litigated to decision twice rather than settling — Realtime Adaptive Streaming (the entity that acquired the portfolio from S.I.SV.EL S.p.A. in 2018) is a monetization vehicle, not an operating company, and it also pursued ex parte reexamination of related patents in its family. (3) Defensive-aggregator presence: the PTAB filings here are by operating companies directly, not Unified Patents — but the district court docket shows the classic Realtime Adaptive Streaming multi-district wave (D. Del., D. Colo., E.D. Tex., D. Mass., C.D. Cal., 17-cv through 19-cv). (4) Patent is dead in the water commercially: Legal Events show lapse on 2025-03-31 ("PATENT EXPIRED FOR FAILURE TO PAY MAINTENANCE FEES," effective 2025-02-21), following a fee reminder mailed 2024-10-14.

Recommended next steps

1. Pull the K1 IPR certificate and both FWDs before doing anything else. They are the only documents that state the claim-level outcome. Sources: PTAB E2E (https://ptacts.uspto.gov/ptacts/) and the docket records at https://www.docketalarm.com/cases/PTAB/IPR2018-01227/Amazon.com_Inc._v._Realtime_Adaptive_Streaming_LLC/ and https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2019-01036. Quote the certificate's cancellation text and the FWDs' "ORDERED that claims ___ ... are unpatentable" language verbatim once you have them.
2. If you are a defendant and claims were canceled, your demand-letter response is: the claims you're being accused of infringing no longer exist in the patent. Link the FWD and the certificate. (I have deliberately not asserted the claim numbers for you, because I could not verify them.)
3. Check the appeal status. Confirm whether CAFC Nos. 21-1518 and 21-2268 are appeals from these two FWDs. If they are and were decided against the patent owner (the 2021-07-21 certificate date is consistent with either no appeal or an affirmance), the invalidity is final and non-appealable. If a FWD was not appealed, that fact alone is telling.
4. Lead with the maintenance-fee lapse. Even setting the IPRs aside, the patent expired 2025-02-21 for non-payment. Any assertion against current conduct is frivolous; only back-damages for pre-expiration infringement within the six-year § 286 window could theoretically be live, and the IPR certificate likely forecloses even that if the asserted claims were canceled.
5. If you need the ODP record to be clean, request a fresh ODP query — the structured block you supplied is stale/empty for these two trial numbers, and any internal reliance on it would have missed the single most important fact about this patent.

Gaps I could not close and will not guess at: (i) FWD claim-level disposition for either trial; (ii) the exact K1 certificate cancellation text; (iii) IPR2019-01036 petition grounds and panel; (iv) whether 21-1518/21-2268 arise from these FWDs.

I'll research the assignment record and ownership chain for US 9,578,298, starting with the reel/frames already surfaced in the patent text.

Let me dig into the assignment correspondents and the Realtime/Sisvel entity structure.

I have enough to reconstruct the chain, though I must flag a correspondent gap. Here is the deliverable.

Sourcing note / correction flag. The USPTO Assignment Center abstract-level fields (conveyance, reel/frame, assignor, assignee, execution and recording dates) are fully reflected below from the Google Patents legal-events record for this patent, which mirrors the same underlying Assignment Center data. The "correspondent of record" field is not exposed in the Google Patents legal-events feed, and I could not retrieve it from any other source within this session. I therefore do not state a correspondent for the three assignment records — see Signal 3, where I explain what I could and could not confirm. This is a gap, not a negative finding.

Inventors

Pattern assessment: No "inventor exodus." All three are career Sisvel personnel. Ballocca's continued rise (to CTO in 2026) confirms the inventors stayed with the original owner — the opposite of the "inventors depart within 12 months of filing, prefiguring a fire-sale" tell. The 2012 assignment (below) was the routine national-stage entry assignment to the applicant, executed ~2 years after the 2010 priority date, not a distress signal.

Original assignee

Discrepancy to flag (contradicting, in part, the earlier section): Google Patents lists the "Original Assignee" as Sisvel SpA and the "application filed by" event as Sisvel SpA. However, the first recorded assignment (Reel 029466/0738) runs from the three inventors directly to Sisvel Technology S.r.l. This is the same inconsistency already noted in the prior section; I am surfacing it again because it matters for the ownership chain — the operative record shows Sisvel Technology S.r.l. as the receiving entity, while the bibliographic "assignee" label says Sisvel SpA. Treat Sisvel Technology S.r.l. → S.I.SV.EL S.p.A. → Realtime Adaptive Streaming LLC as the chain of record.

• Primary line of business: Sisvel Technology S.r.l. was the in-house R&D arm of the Sisvel Group — a patent-licensing / patent-pool administrator. It developed the "Tile Format" backward-compatible 3D-TV packing scheme (Ballocca, D'Amato, Grangetto, Lucenteforte — IEEE ICME 2011), which is the subject matter family of this patent.
• Product embodying the claims? No commercial product. Sisvel Technology built demos and ran 3D-TV broadcast trials, but its business model is licensing (DVB-H pool, VP9/AV1 pools), not selling devices. This matters for the NPE analysis: even the original assignee is a licensing shop, not a product company.
• Current status: The Sisvel Group (S.I.SV.EL Società Italiana per lo Sviluppo dell'Elettronica S.p.A., Luxembourg/Italy) remains an operating licensing business — it is a live pool administrator (recent news includes AI-driven MPEG-5 EVC work and a 2026 CTO appointment). The specific U.S. right, however, left the Sisvel Group in 2017 (see below) and the patent lapsed 2025-02-21 for non-payment of maintenance fees.

Assignment timeline

Three recorded assignments. Reel/frame identifiers are from the Google Patents legal-events feed (which mirrors USPTO Assignment Center).

• 2012-11-19 (executed) / recorded 2012-12-13 — Reel 029466/0738

  • Conveyance: Assignment
  • Assignor: Ballocca, Giovanni; D'Amato, Paolo; Celia, Saverio (the inventors)
  • Assignee: Sisvel Technology S.r.l. (Italy)
  • Correspondent: not disclosed in accessible sources — see Signal 3. No recurrence determination possible.
  • Context: Ordinary inventor-to-applicant assignment accompanying U.S. national-stage entry (App. 13/704,217); standard prosecution chain-of-title, not a distress transfer.

• 2013-06-28 (executed) / recorded 2014-07-29 — Reel 033415/0144

  • Conveyance: Assignment (NUNC PRO TUNC) — recorded expressly as retroactive
  • Assignor: Sisvel Technology S.r.l.
  • Assignee: S.I.SV.EL Società Italiana per lo Sviluppo dell'Elettronica S.p.A.
  • Correspondent: not disclosed in accessible sources — see Signal 3.
  • Context: Internal corporate reorganization within the Sisvel Group — the nunc pro tunc character and the ~13-month gap between execution and recording (plus a later certificate of correction, 2017-08-29) indicate a housekeeping/retroactive clean-up of intra-group title, not an arm's-length sale.

• 2017-08-29 (executed) / recorded 2018-11-16 — Reel 047531/0016

  • Conveyance: Assignment
  • Assignor: S.I.SV.EL Società Italiana per lo Sviluppo dell'Elettronica S.p.A.
  • Assignee: Realtime Adaptive Streaming LLC (Texas; principal place of business 1828 E.S.E. Loop 323, Tyler, Texas 75701)
  • Correspondent: not disclosed in accessible sources — see Signal 3.
  • Context: Transfer-to-asserter. The patent exits the Sisvel Group to a litigation entity; the execution date (2017-08-29) is only ~3 months before the first suits naming the '298 patent (fall 2017). Note this right is an outlier in Realtime's portfolio — Realtime's core Fallon-family patents (the '046/'535/'610/'477 family) came from Realtime Data LLC → Realtime Adaptive Streaming at Reel 041573/0683 (2017), whereas the '298 came straight from S.I.SV.EL, i.e., a different provenance ("Realtime Data LLC d/b/a IXO" is the well-known parent brand).

No further assignments recorded after 2018-11-16. Realtime Adaptive Streaming LLC remains the assignee of record through lapse.

Timeline diagram

timeline
    title Ownership of US 9578298
    2010 : Priority date June 28
         : Invented by Sisvel Technology team
    2011 : PCT application filed June 27
    2012 : Inventors assign to Sisvel Technology
    2013 : Nunc pro tunc transfer to S I SV EL
    2014 : Sisvel internal reassignment recorded
    2017 : Patent issued February 21
         : Assigned to Realtime Adaptive Streaming
         : First suits naming the patent filed
    2018 : Realtime assignment recorded
    2025 : Lapsed for unpaid maintenance fee

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT. The patent moved from the Sisvel Group (a licensing operating business) to Realtime Adaptive Streaming LLC, a Texas LLC with a Tyler, Texas address and no products in commerce, by Reel 047531/0016 (executed 2017-08-29, recorded 2018-11-16). Realtime's own PTAB filing states its principal place of business as 1828 E.S.E. Loop 323, Tyler, Texas 75701 (IPR2018-01187 Power of Attorney). A single-purpose, assertion-only LLC is evidenced by its docket: dozens of suits (Apple, Google, Netflix, Samsung, Sony, Cisco, AMD, Intel, Broadcom, Adobe, Avaya, Cox, EchoStar/DISH) — not product sales.

2. Known asserter in the chain — PRESENT. Realtime Adaptive Streaming LLC is a documented high-frequency patent plaintiff (RPX litigation documents; Unified Patents PTAB records; Federal Circuit Realtime Adaptive Streaming L.L.C. v. Sling TV, L.L.C., 113 F.4th 1348 (Fed. Cir. 2024), and Nos. 21-1518 / 21-2268). It is not one of the "classic" listed entities (Acacia, Marathon, IV, Wi-LAN, etc.), but it is squarely a repeat NPE surfaced by both RPX and Unified Patents. Affiliate Realtime Data LLC (d/b/a IXO) is the better-known entity in the same orbit.

3. Repeat correspondent across the chain — UNCLEAR / not confirmable. The correspondent-of-record field on the three assignment abstracts (Reels 029466/0738, 033415/0144, 047531/0016) is not exposed in the Google Patents legal-events feed, and I could not retrieve it from a second source this session, so I cannot test for a recurring recording attorney. What I can report (litigation counsel — a related but distinct data point, not the assignment correspondent): Realtime's PTAB/appellate counsel is Russ August & Kabat (12424 Wilshire Blvd., Los Angeles — Neil Rubin and Kent Shum, Reg. No. 61,117), with Eric B. Fenster, LLC listed as plaintiff's district-court counsel. Do not treat these as assignment correspondents. To close this signal, run the patent-number search on the Assignment Center and read the "Correspondent" line for each reel.

   • Assignment Center search to run: https://assignmentcenter.uspto.gov/ (patent no. 9578298; also check the alternative index https://assignment.uspto.gov/patent/index.html).

4. Cascading transfers — PARTIALLY PRESENT. Three recordings across two ownership regimes: two intra-Sisvel links (2012, 2013/2014) then one exit to Realtime (2017/2018). The intra-group pair is not the classic "<24-month chained-LLC" pattern (both are Sisvel entities), but the overall sequence — inventors → operating affiliate → holding company → asserter — is a three-hop chain. Cite Reels 029466/0738, 033415/0144, 047531/0016.

5. Pre-litigation transfer — PRESENT. Assignment executed 2017-08-29 (Reel 047531/0016); the '298 patent is named in suits filed in the fall 2017 window (e.g., D. Del. Netflix/Haivision, C.D. Cal., D. Colo. 1:17-cv-02869 v. Apple, and E.D. Tex. — all reflected in the AO 120 § 290 filings for Realtime). That places the transfer within ~3 months of first assertion — squarely inside the 6-month window that signals the chain was arranged to enable suit. (Recording lagged execution by ~14 months, but execution controls.)

6. Bankruptcy fire-sale — NOT PRESENT. No Chapter 7/11 for any assignor; the Sisvel Group is solvent and operating. The exits were a licensing-company portfolio divestiture and an intra-group reorganization, not insolvency sales.

7. Privateering — NOT PRESENT (does not fit the fact pattern). Sisvel is itself a licensing/pool entity, not a product-selling operating company asserting against competitors through a proxy. The 2017 transfer is best characterized as a portfolio divestiture by one licensing entity to another, not classic operating-company privateering. (No SEC 10-K/8-K disclosure is available because neither Sisvel nor Realtime is a U.S. public operating company.)

8. Defensive aggregator — NOT PRESENT. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at an asserting NPE, the inverse of neutralization. (Separately, the Sisvel–RPX VP9/AV1 transaction involves different Sisvel assets and is not this patent.)

Verdict

NPE — high confidence.

Driven by two strong signals plus one supporting signal on the specific recorded chain: (i) a transfer-to-asserter from S.I.SV.EL S.p.A. to the single-purpose Texas LLC Realtime Adaptive Streaming LLC at Reel 047531/0016 (executed 2017-08-29, recorded 2018-11-16); (ii) an identified, high-frequency NPE as current assignee of record — Realtime Adaptive Streaming LLC appears as plaintiff across dozens of suits (Apple, Google, Netflix, Samsung, Sony, Cisco, AMD, Intel) and in Federal Circuit appeals, with a Tyler, Texas address and no products; and, supporting, (iii) a pre-litigation transfer executed within ~3 months of the first '298 suits. The patent never embodied a sold product — even the original assignee (Sisvel) was a licensing business — and it lapsed 2025-02-21 for non-payment of maintenance fees while still held by Realtime.

Caveat on confidence: Signal 3 (repeat correspondent) could not be evaluated because the assignment correspondents were not retrievable in this session. If the correspondents on Reels 029466/0738, 033415/0144, and 047531/0016 reveal a single recurring recording attorney across the Sisvel and Realtime links, that would further reinforce the verdict; if they do not, the verdict is unchanged on the strength of Signals 1, 2, and 5.

Verify at: https://assignmentcenter.uspto.gov/ (search patent 9578298) and the mirror at https://assignment.uspto.gov/patent/index.html.

I'll research the specific prior-art citations for US 9,578,298, grounding each reference and its § 102 relevance.

Let me pull details on the individual cited references to ground the anticipation analysis.

Let me gather details on the remaining key citations.

Prior-Art Analysis — US 9,578,298 B2 ("Method for decoding 2D-compatible stereoscopic video flows")

Scope, method, and identifier handling

I did not auto-correct any identifier. I searched the literal strings 9578298 / 9,578,298. This is Sisvel's 2D-compatible-stereo-decoding patent; it is not any other "’298" patent (e.g., Amgen's erythropoietin patent or U.S. 6,164,298) — those hits were excluded.

On the "USPTO database" instruction: I could not directly query the USPTO PatentCenter/Patent Full-Text databases from this session. Instead I worked from (a) the authoritative full text you provided (which reproduces Google Patents' "Patent Citations" list — the references of record on the face of the patent), and (b) public mirrors (Google Patents, FreePatentsOnline, Justia, PatentScope/EPO, USPTO PTAB). Where a bibliographic detail could not be independently confirmed, I flag it rather than assert it. This is a gap, not a confirmed negative.

Legal framework. The application was filed 2011-06-27 (pre-AIA, since before 2013-03-16), claiming priority to IT TO2010A000549 filed 2010-06-28. Accordingly pre-AIA 35 U.S.C. § 102(a)/(b)/(e) governs. A reference anticipates a claim only if it discloses every limitation as arranged in the claim; otherwise the reference is at most § 103 material. I apply that standard strictly below.

Important structural point about the reference set. Of the 17 patent citations of record, several are cited only in the Background as descriptions of known 2D-compatible 3D coding techniques (depth maps, multiview/MVC). Those are § 102(b)-type printed publications but, as shown below, cannot anticipate because they lack the composite-frame/one-image-extraction architecture. The genuinely probative references cluster around composite-frame signaling (Thomson, Sasaki, ETRI, Samsung, 3Dswitch) and 2D/3D switching (Philips, RealD).

⚠️ Distinction to keep clear: the references of record (below) are not the same as the art actually used in the IPRs. IPR2019-01036 (Google/YouTube) relied on US 2010/0225645 A1 (Suh) in view of US 2006/0015919 A1 (Karppinen) — neither is on the face of the '298 patent. I treat those separately in § 7.

1. Patent citations of record (the 17 references)

Dates are given as priority (earliest) / publication. § 102 basis is provisional and keyed to the 2010-06-28 priority.

Also cited in the family (not on the U.S. face): WO 2006/137000 A1 — Koninklijke Philips, "Combined exchange of image and related data" (2005-06-23 / 2006-12-28) — listed under "Family Cites Families." Relevant § 103 context for coupling image + associated data (signaling).

2. Ranked assessment: the most relevant prior art

1. WO 2008/127676 A2 (Thomson Licensing — "Tiling in video encoding and decoding") — the single closest reference to claim 1: multi-picture-in-one-picture + combination metadata + decode/provide a representation of at least one of the pictures + output. Its only arguable gap is the literal "decoding only that part" of claim 1. Recommend this as the lead § 102/§ 103 reference.
2. US 2010/0245548 A1 (Sasaki / Panasonic) — the reference the examiner used for "decoding only that part" (2D device decodes only the left-view). Strongest on that specific limitation (subject to the § 102(e) date check and the extent-skipping distinction).
3. US 2012/0105583 A1 (Suh / LG) and WO 2008/054100 / JP 2010-508752 (ETRI) — strongest on the metadata/signaling limitations (claims 3–10).
4. WO 2009/077929 A1 (Philips) and US 2010/0026783 A1 (RealD) — strongest on 2D/3D mode selection/user switching (claims 13–14) and in-frame signaling (claim 6), respectively.
5. US 2008/0303893 A1 (Samsung) and WO 2010/046739 / US 2011/0234753 (3Dswitch) — supporting § 103 art for header/in-frame metadata and packing-format description.

Bottom line under § 102: On my reading, no single cited reference on the face of the '298 patent discloses every limitation of independent claim 1 (and thus of claims 11/19) as arranged in the claim, primarily because the record references either (a) decode the entire combined frame and then select a view (Thomson, Samsung), or (b) obtain "one view only" by transport/stream-level skipping rather than by region-selective decoding of a composite frame (Sasaki). Anticipation is therefore reference- and claim-construction-dependent, and the realistic attack vector is § 103 (or the IPR art in § 7).

3. What each dependent claim pulls from the cited art

• Claim 2 (image undivided): met by any of the packed-frame references (Thomson; Sharp EP 1 524 859) if the frame is a side-by-side/top-bottom packing. § 103.
• Claims 3, 7–10 (coordinates of an area / polygon(s) / vertices): best supported by WO 2008/054100 / JP 2010-508752 (ETRI) and WO 2008/127676 (Thomson) (combination "information" identifying where the pictures sit). § 103.
• Claims 4, 5, 6 (metadata in H.264 SEI / Service Information / within the composite frame): WO 2008/127676 (SEI-level high-level syntax is expressly discussed in the tiling family); US 2008/0303893 (Samsung) and US 2010/0026783 (RealD) for in-frame/header carriage. § 103.
• Claims 13, 14 (user-command / remote-control activation of the 2D decode): WO 2009/077929 (Philips) is the closest (user-controllable playback setting; 2D/3D switching). § 103.

4. Non-patent citations of record (context, not prior art per se)

• International Search Report, PCT/IB2011/052810, 2011-11-17.
• Written Opinion, PCT/IB2011/052810, 2011-11-17.
• European Office Action, EP 11744071.9, 2015-01-06.
• Japanese Office Action, JP 2013-517623, 2015-08-25.

The ISR/Written Opinion identify the examiner's original art selection and are the best guide to the intended § 102/§ 103 theory of the record.

5. Additional highly relevant art used against the patent (not face citations)

These are not on the patent's face but are the operative prior art in the validity challenges, and they are more dangerous than most of the record references:

• US 2010/0225645 A1 (Suh) — primary reference in IPR2019-01036 (Google/YouTube), combined with:
• US 2006/0015919 A1 (Karppinen) — secondary reference.

The dispute in that IPR centered on the claim-1 limitation "decoding only that part of the composite frame (FC) which contains said one image to be displayed." Also, in IPR2018-01227, the examiner/Board discussion paired Kim and Cho with Sasaki for that same limitation. If you are assessing validity, these IPR references should be added to the citation list — they post-date the face citations and are the art the patent actually had to survive.

6. Contradiction / accuracy flags (per your cross-reference rule)

• No material contradiction between the previously generated summary and this analysis. The prior section's claim count (19), issue date (2017-02-21), and lapse (maintenance fee lapsed 2025-03-31, effective 2025-02-21) are consistent with the authoritative text.
• One item requiring verification, not correction: the prior summary flagged an unresolved discrepancy over the IPR2018-01227 petitioner (Amazon vs. "prior Netflix IPR"). The '298 legal-events record shows the IPR2018-01227 opponent as Amazon.com, Inc. / Amazon Digital Services LLC, which resolves the petitioner identity for that docket; the "Netflix" reference in the later Google IPR appears to be a drafting artifact in the patent-owner's paper, not a separate proceeding.
• Date/§ 102(e) uncertainties (explicit gaps, not assertions): the § 102(e) eligibility of US 2010/0245548 (Sasaki) and US 2012/0105583 (Suh/LG) depends on their U.S. filing dates relative to 2010-06-28, which I could not confirm without direct USPTO access. Confirm in PatentCenter before relying on either as § 102(e) art.

7. Recommended next steps for a definitive record

1. Pull the "References Cited" page of the issued patent and the PTO-892/1449 forms from USPTO PatentCenter (App. 13/704,217) to confirm whether any of the 17 were applicant-supplied vs. examiner-cited (this affects the § 103 "teaching away"/motivation narrative).
2. Verify the § 102(e) dates for US 2010/0245548 and US 2012/0105583.
3. Retrieve the full text of WO 2008/127676 (Thomson) claim set and the Sasaki 2D-playback-decode passage to finalize the § 102 vs. § 103 line for claim 1 — these two references determine the outcome.
4. Add the IPR art (US 2010/0225645 Suh; US 2006/0015919 Karppinen) to the prior-art chart even though neither is a face citation.

Key sources: https://patents.google.com/patent/US9578298/en (full text, claims, citations, legal events) · https://patents.google.com/patent/WO2008127676A3/en · https://www.freepatentsonline.com/[9232235](/patent/9232235).html · https://patents.google.com/patent/WO2009077929A1/en · https://patents.google.com/patent/US20100026783A1 · https://uspto.report/patent/grant/[8149267](/patent/8149267) · https://portal.unifiedpatents.com/ptab/case/IPR2019-01036 and .../IPR2018-01227.

Everything above is a technical prior-art assessment, not a legal opinion on validity, infringement, or patentability.

I'll ground my obviousness analysis in the prior-art references actually cited on this patent's face, plus the IPR record. Let me pull authoritative content on the key references.

Strong grounding so far. Let me now pull the remaining key cited references, including the Philips 3D-mode reference (relevant to the user-command claims) and the Sharp 3D video coding references.

Obviousness Analysis — U.S. Patent No. 9,578,298 B2

"Method for decoding 2D-compatible stereoscopic video flows"
Priority: 2010-06-28 (IT TO2010A000549) · PCT/IB2011/052810 filed 2011-06-27 · Issued 2017-02-21 · 19 claims

0. Scope, method, and caveats

Per the operating rules, I interpret patent/publication identifiers literally and do not auto-correct them. The analysis below is built from (a) the full text of US 9,578,298 and (b) the Prior Art section of that same page — i.e. the 15 references listed as "Cited By/Citations (15)" and the fuller "Patent Citations (17)" list, plus the 4 non-patent citations. Where I rely on the IPR record rather than the face of the patent, I say so explicitly.

Two honesty flags before the merits:

• I could retrieve and read substantive text for WO 2008/054100 A1, US 2010/0026783 A1, WO 2008/127676 A2, US 2010/0245548 A1, and WO 2009/077929 A1. For EP 1 524 859 A2 / JP 2004-240469 A / JP 2005-124200 A (Sharp), WO 2007/064159 A1 (ETRI), US 2008/0303893 A1 (Samsung), WO 2010/046739 A1 (3Dswitch), US 2012/0105583 A1 (Suh), and the 3Dswitch and Harman/DDD documents, I have only the bibliographic data on the patent page plus what limited text the searches returned. I do not assert specific teachings for those beyond their titles/abstracts, and I flag them as lower-confidence mappings.
• The earlier-generated sections of this analysis cover the bibliographic record, claim construction exposure (Adobe claim-construction order), and litigation/PTAB history. I do not repeat that; I build on it.

1. The legal frame and the real battleground

Under Graham v. John Deere and KSR Int'l v. Teleflex, the question is whether the claimed subject matter as a whole would have been obvious to a PHOSITA at the 2010-06-28 priority date, considering (i) the scope/content of the prior art, (ii) differences, (iii) level of ordinary skill, and (iv) objective indicia.

Level of ordinary skill: a bachelor's degree in EE/CS (or equivalent) plus 2–4 years in digital video compression/MPEG transport, including frame-packing ("stereoplexing") and broadcast/CE receiver design. This is a mature, standards-driven field (MPEG-2 MVP, H.264/MVC, DVB, Blu-ray 3D), which matters under KSR: in a field where market demand and standardized interop dictate the design space, "the combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results."

What is not in dispute — the admitted prior art. The '298 specification concedes essentially the entire factual predicate of claim 1:

• Frame packing (side-by-side, top-bottom) into a single composite frame: admitted as conventional (FIGS. 1a–1b).
• Metadata describing the packing format, including where it is carried ("SEI messages of the H.264 coding," "Service Information," or inside the frame per the applicant's own IT2008MO000267): admitted as "a common practice."
• The "explicit vs. implicit" metadata dichotomy claimed in claim 1 (geometry or packing type): admitted, and expressly attributed to the prior-art signalling convention.
• 2D-compatibility as the stated design goal: admitted as the central industry requirement.

What matters, then, is only the residual limitation: claim 1's step "decoding only that part of the composite frame (FC) which contains said one image to be displayed," plus the dependent-claim geometry details (claims 7–10) and the user-triggered switching (claims 13–14). If the cited art teaches selective/partial decode of a sub-region of a composite frame, the case for § 103 is strong across the board. And it does.

Note also that the patent's own specification supplies the motivation that defeats any nonobviousness argument on this point:

"those parts of the composite frame which are not contained in the 2D display window … are not used by the visualization apparatus and, in principle, may even not be decoded." ('298 spec.)

That is an express admission that skipping decode of the unused view is an available and sensible design choice — powerful 103 ammunition, since a reference need not teach what the applicant concedes was known.

2. Claim 1 — limitation-by-limitation map against the cited art

Observation on scope: claim 1's step (d) is arguably anticipated by WO 2008/127676 A2 read alone — the Thomson claims literally recite decoding only "at least one of the multiple pictures" from a combined picture, which is the 2D-extraction use case. Where single-reference anticipation is contested (because Thomson's "decoding" may occur at a picture-composition level rather than at the entropy-decode level), the limitation is at minimum obvious in view of a second reference; hence the § 103 framing below remains the safe analytical path.

3. Primary § 103 grounds

Ground 1 — WO 2008/054100 A1 (ETRI) + WO 2008/127676 A2 (Thomson) ⇒ claims 1, 3, 5, 6, 11, 15–19

Coverage. ETRI supplies the composite-frame + packing-format-metadata architecture (limitations a–b) including the explicit "packing type" signalling codes (side-by-side / top-down) and the express 2D-legacy-compatibility and user-selectable 2D/3D mode rationale. Thomson supplies the partial-decode / partial-output mechanism (limitations c–e): combining information that locates each constituent picture within a single coded picture, decoding "at least one" of them, and outputting the decoded representation of the selected one.

Motivation to combine (articulable rationale under KSR):

1. Same field, same problem, same solution family. Both address backward-compatible carriage and selective presentation of stereo/multi-view content in a single coded picture. KSR holds that combining references directed to the same field to solve the same problem is obvious.
2. Complementary, not overlapping. ETRI answers "how do I signal/describe the packing in the bitstream?" Thomson answers "how do I decode and extract only one of the packed pictures?" One reference supplies what the other lacks; the combination is the natural pairing a PHOSITA would reach for when implementing a 2D-compatible receiver.
3. Predictable result. The combination achieves nothing more than the sum: a receiver that parses packing metadata and decodes the corresponding sub-region. No unexpected property, no new interoperability problem; accordingly the KSR "predictable results" rationale applies directly.
4. Design incentive exists in the references themselves. ETRI's stated object is enabling "a user having an existing player [to] view a 2D image," and Thomson's object is user-selectable view extraction. Both point the PHOSITA toward the claimed output stage.

Where the combination is strongest: limitations (a), (b) (both the "geometry" and "packing type" alternatives are independently met), and the output-frame generation of (e).

Where a secondary reference tightens the case: adding WO 2009/077929 A1 (Philips) — which claims a stored, user-settable 2D/3D playback setting and processing of "either the primary video stream or the primary video stream and the additional information stream, in accordance with the playback setting" — supplies limitation (d) in mode-selection terms (only the 2D-necessary stream/portion is processed), reinforcing the "decode only that part" step.

Ground 2 — US 2010/0026783 A1 (RealD) + WO 2008/127676 A2 (Thomson) ⇒ claims 1, 3, 6, 11, 12, 15–19

Coverage. RealD teaches: receiving 3-D video in a "transportable format (e.g., side-by-side)," embedding a machine-readable tag in the frame to signal 3-D content, detecting the tag at the receiver, and switching the receiver into/operating in 2-D (or 3-D) mode, including "allow[ing] bottom lines to pass through unaffected" and "replac[ing] tag pixels with black pixels." That is: metadata carried within the composite frame (claim 6), detection, identification, and mode-dependent processing. The remaining piece — using the layout information to select and decode only the desired view region — comes from Thomson, exactly as in Ground 1.

Motivation: identical to Ground 1, plus a strong KSR "known technique" rationale — RealD and Thomson both solve "deliver 3-D content through 2-D infrastructure without breaking legacy devices," and a PHOSITA implementing RealD's automatic 3-D/2-D detection would plainly look to Thomson for the mechanics of extracting the selected view rather than re-deriving them.

Both RealD and Thomson predate the priority date on their faces (RealD published 2010-02-04; Thomson published 2008-10-23), so neither raises a § 102(e)-date complication for claim 1.

Ground 3 — The IPR ground of record: US 2010/0225645 A1 (Suh) + US 2006/0015919 A1 (Karppinen) ⇒ claims 1–4, 7–11, 13

This is from the IPR record, not the "Prior Art" face of the '298 page, so I present it as supplementary and flag the provenance. Per the earlier-generated litigations summary, IPR2019-01036 (Google LLC / YouTube LLC, filed 2019-05-06) challenged claims 1–4, 7–11, and 13 as obvious over US 2010/0225645 A1 (Suh) in view of US 2006/0015919 A1 (Karppinen), with the dispute focused precisely on "decoding only that part of the composite frame (FC) which contains said one image to be displayed." The mapping is structurally the same as Grounds 1–2: Suh supplies composite-frame/packing-format signalling for stereoscopic video; Karppinen supplies the selective-region decoding/processing component. The Board's institution on those claims is itself evidence that a PHOSITA would have combined them with a reasonable expectation of success.

⚠️ Cross-reference flag: the earlier "Patent summary" section records that a single IPR certificate (kind code K1) issued 2021-07-21 covering both IPR2018-01227 and IPR2019-01036, and that I could not confirm claim-by-claim which claims were cancelled vs. upheld. I preserve that uncertainty here — I do not assert that any particular claim was held unpatentable.

Ground 4 — 2D-compatibility architecture generally: WO 2010/046739 A1 (3Dswitch) + WO 2008/054100 A1 (ETRI)

Lower confidence. Both are titled/described as systems for encoding and decoding stereoscopic images with 2D compatibility. If 3Dswitch discloses region/geometry-based extraction of one view from a packed stereoscopic image (consistent with its title and the family's subject matter), it would provide limitation (c–d), which combined with ETRI's packing-type metadata (a–b) covers claim 1. I do not have full text for 3Dswitch and therefore present this ground as a hypothesis to be verified against the document, not a grounded mapping. Notably, 3Dswitch's U.S. sibling US 2011/0234753 A1 appears in the patent's own citation list, so the Examiner considered this family.

Ground 5 — Dependent-claim grounds

4. The device claims (11–19)

Claim 11 is the apparatus mirror of claim 1 (receiver + scaler + decoder), and claim 19 is a video playback device comprising the claim-11 device. Two structural points help the § 103 case:

1. The "scaler" element is expressly admitted as conventional. The '298 spec states that "the subsequent resizing of the selected section can be delegated to the scaler of the decoder 212 or … to the one of the display device 222," applying "suitable interpolation algorithms." Philips (WO 2009/077929 A1) and RealD both describe display-side processing units/rendering stages performing exactly the resolution/format restoration role. Mapping a known scaler to a known decoder is a KSR "arrangement of elements" case.
2. Claim 19 adds no new functionality; it is the standard packaging of the claim-11 device into a playback device (ETRI's "stereoscopic player structure," RealD's decoder/receiver, Philips' "playback device"). Under KSR, mere recitation of a new end-use environment for a known device is not patentably distinct where the environment is the device's evident application.

Ground for 11–19: Grounds 1 or 2 above, applied to the apparatus, with Philips (WO 2009/077929 A1) supplying the playback-device/2D-3D-setting structure and RealD supplying the receiver+scaler topology.

5. Articulated motivation-to-combine summary (the § 103 "why")

A PHOSITA at 2010-06-28 had, and the cited references supply, five independent rationales:

1. Common problem, common field. Every reference is directed to 2D-compatible carriage of stereoscopic content in a single coded video picture for broadcast/BD/streaming. Same problem ⇒ combination is obvious (KSR).
2. Complementary teachings with no overlap. Metadata/signalling (ETRI; RealD tag) + selective extraction/decode (Thomson) are the two natural halves of one receiver design.
3. Predictable, incremental result. No new technical effect arises; each reference performs its known function. KSR "predictable results" rationale.
4. Explicit design incentive in the references. ETRI and RealD both state the objective of automatic 2D/3D identification and mode switching; a PHOSITA would not need hindsight to reach the claimed output.
5. Admitted prior art and admitted design choice. The '298 spec admits packing-format metadata carriage (SEI/SI) and expressly contemplates not decoding the unused region — removing the only arguably novel step from the field of the non-obvious.

6. Counterarguments the patent owner could raise (and the response)

To be balanced, and because the earlier sections flagged genuine uncertainty in the record:

• "Only that part … decode" is a decoding-architecture limitation, not merely a presentation choice. The owner's best argument is that the cited references decode all pictures of the combined frame and then extract the desired one for display, so the references do not teach entropy-level partial decode. Response: Thomson's claims recite decoding "at least one of the multiple pictures" (not all), and the specification's admission that unused parts "may even not be decoded" is an express concession that the step was within the ordinary skill level. The IPR ground of record (Ground 3) targeted exactly this limitation, which suggests the Board found the argument, at minimum, not dispositive at institution.
• Claims 8, 9, 16, 17 (polygon / plurality of polygons). These require express vertex-of-polygon signalling. The cited face of the '298 patent does not give me a clean single-reference disclosure of polygon (as opposed to rectangular-area) vertex coordinates. If these claims were not among those cancelled by the 2021 IPR certificate, they are the most defensible. Flagged as a genuine weakness in my mapping, not a conclusion.
• Claim 2 ("undivided") could be argued to distinguish over multi-tile/quad-tile schemes (RealD "stereoplexing," 3Dswitch) in which a single view is split across areas. Response: side-by-side and top-bottom packing preserves each view as one undivided half-frame; ETRI and RealD both disclose those formats.
• Secondary considerations. The litigation record (Google Patents "Family has litigation": D. Del., C.D. Cal., D. Colo., E.D. Tex., D. Mass.) could support commercial-success/industry-adoption narrative. But the earlier-generated sections establish no final infringement judgment or damages award on the '298 patent; the actions were dismissed, stayed, settled, or superseded by the IPRs, and the patent lapsed for non-payment of maintenance fees effective 2025-02-21. The nexus and timing of any such indicia would need independent proof — I do not treat them as established.

7. Bottom line

• Claims 1, 3–7, 10–16, 18, 19 are, in my assessment, likely obvious under § 103 over WO 2008/054100 A1 (ETRI) in view of WO 2008/127676 A2 (Thomson), and independently over US 2010/0026783 A1 (RealD) in view of WO 2008/127676 A2 (Thomson), with WO 2009/077929 A1 (Philips) supplying the user-command/mode-selection features of claims 13–14 and the playback-device structure of claim 19. Confirmed by the independent IPR ground US 2010/0225645 A1 (Suh) + US 2006/0015919 A1 (Karppinen).
• Claims 8, 9, 16, 17 (polygon / plurality-of-polygon vertex coordinate signalling) are the strongest candidates to survive, because I could not verify a clean explicit polygon-vertex disclosure on the cited face of the patent. 3Dswitch (WO 2010/046739 A1) and the applicant's own IT2008MO000267 are the most likely sources to check.
• Confidence: High for Grounds 1 and 2 and the dependent-claim mappings in the table's "very high / high confidence" rows; moderate-to-low for Grounds 4–5 and for the polygon claims. Per the operating rules, I have not asserted specific teachings for references whose full text I could not retrieve, and I have not auto-corrected any identifier (e.g., I treat the "Citations (15)" vs. "Patent Citations (17)" counts as the page states them, and I did not reconcile the petitioner discrepancy in IPR2018-01227).

Key source URLs: https://patents.google.com/patent/US9578298/en · https://patents.google.com/patent/WO2008054100A1/en · https://patents.google.com/patent/US20100026783 · https://patents.google.com/patent/WO2008127676A2/en · https://patents.google.com/patent/WO2009077929A1/en · https://patents.google.com/patent/US20100245548A1 (page reference) · https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1525803](/patent/1525803) (IPR2019-01036 petition referencing Suh '645 and Karppinen '919).