US 11336832

I'll search for authoritative records on this patent, including the CAFC docket mentioned in the patent metadata.

Let me get the PTAB final written decision and the exact independent claim text.

Let me pin down the independent claim 11 text and the PTAB final written decision details.

US 11,336,832 B1 — Verification Summary

Search scope note: I searched for the exact identifier 11336832 / US11336832B1 only. Results for other GoPro patents in the same litigation family (e.g., 10,015,413; 10,529,052; 10,574,894; 10,958,840; D789,435) were excluded from the discussion below except where they are part of the same proceeding.

1. Bibliographic data (from the authoritative patent text)

Abstract (verbatim): "A video may be captured by an image capture device in motion. A horizon-leveled view of the video may be generated by providing a punchout of the video. The punchout of the video may compensate for rotation of the image capture device during capture of the video. The placement of the punchout of the video may be changed based on different rotational positions of to provide a view in which a horizon depicted within the video is leveled."

Family members noted in the record: US17/744,949 → US12302000B2, and US18/750,643 → US20240348927A1.

Security interests recorded: JPMorgan Chase N.A. as administrative agent (2020-10-19), released 2021-01-25; Farallon Capital Management L.L.C. and Wells Fargo Bank, N.A. as agents (both 2025-08-04).

2. Independent claims — plain-language overview

The patent has at least 20 claims. Claim 1 (system) and claim 11 (method) are the independent claims; this is confirmed by GoPro's Patent Owner Response in IPR2024-01052, which refers to "independent claims 1 and 11 of the '832 patent."

Claim 1 — "A system for horizon leveling videos" (text confirmed verbatim from ITC Inv. No. 337-TA-1400 Initial Determination, quoting JX-0009):

A system with one or more programmed physical processors that:

1. Obtain video information defining a video captured by an image capture device over a capture duration, where the video content has a progress length, contains visual content captured at different moments, is viewable as a function of progress through the progress length, and has a field of view.
2. Obtain rotational position information for the video characterizing rotational positions of the image capture device as a function of progress through the capture duration (i.e., frame-by-frame orientation data).
3. Determine a viewing window for the visual content as a function of progress through the progress length, based on those rotational positions. The window defines which extents of the visual content go into horizon-leveled visual content. Determining the window includes determining a placement of the window within the field of view as a function of progress (i.e., the crop is re-positioned/re-oriented over time, not fixed once).
4. Generate horizon-leveled visual content based on the viewing window, comprising a punchout of the extents defined by the window, where including those extents "effectuates horizon leveling."

In plain terms: record gyro/accelerometer-based roll data per frame; slide and rotate a crop window frame-by-frame so the crop is level; output the cropped, level video.

Claim 11 — independent method claim. I was unable to retrieve the verbatim text of claim 11 from the sources reviewed. Based on (a) GoPro's POR characterizing claims 1 and 11 as the independent claims, (b) the petition challenging claims 1, 2, 7–12 and 17–20 together, and (c) the parallel dependent-claim structure (13–16 mirror 3–6; 17–20 mirror 7–10), claim 11 is a method claim reciting steps that parallel claim 1's elements (obtain video information → obtain rotational position information → determine a viewing window as a function of progress → generate horizon-leveled visual content comprising a punchout). Treat this as an inference, not a verified quotation.

Dependent-claim architecture (for context):

• Claims 3–6 / 13–16: visual content contains a distortion (barrel or pincushion; from a wide-FOV optical element or applied to non-distorted content) so a straight line appears curved, reducing the visual impact of an off-axis horizon.
• Claims 7–10 / 17–20: the viewing window determination further includes size (7/17, changing to simulate zoom — 8/18), rotation (9/19), and location/framing (10/20).

3. Claim construction of note

In IPR2024-01052, the Board construed the "determining a viewing window … as a function of progress through the progress length" limitation as requiring a determination of a viewing window for different times of the video — i.e., a single one-time determination applied to every frame of the entire video would not meet the limitation. This construction is the crux of the dispute with the Thomason reference (which the Board found disclosed only a one-time orientation measurement).

4. Litigation and administrative proceedings (searched)

a) PTAB — IPR2024-01052, Arashi Vision Inc. (d/b/a Insta360) v. GoPro, Inc.

• Petition filed 2024-06-20 by Arashi Vision Inc. (d/b/a Insta360) as Petitioner; GoPro as Patent Owner.
• Institution decision 2025-01-07.
• Final Written Decision 2025-12-02 — outcome recorded as "Unpatentable." Reported breakdown: claims 1, 2, 7–12, and 17–20 held unpatentable under 35 U.S.C. § 103; claims 3–6 and 13–16 not shown to be unpatentable (the distortion claims survived on the grounds presented).
• Status: Final Written Decision – Appealed. GoPro also filed a Request for Director Review challenging the unpatentability findings on claims 7–8 and 17–18.
• Petitioner's grounds relied on Thomason (Ex. 1004) and Pacurariu, among others.

b) CAFC — Docket No. 26-1859, Arashi Vision US LLC v. GoPro Inc

• Appeal from the PTAB, originating case IPR2024-01052.
• Reported filed 05/21/2026; status pending judge assignment; no outcome recorded.
• URL surfaced in the patent record: https://portal.unifiedpatents.com/litigation/Court%20of%20Appeals%20for%20the%20Federal%20Circuit/case/26-1859

c) ITC — Inv. No. 337-TA-1400, Certain Cameras, Camera Systems, and Accessories Used Therewith

• GoPro complaint filed 2024-03-29 asserting, inter alia, claims 1–10 of U.S. Patent No. 11,336,832. Respondents: Arashi Vision Inc. d/b/a Insta360 and Arashi Vision (U.S.) LLC d/b/a Insta360.
• ALJ Doris Johnson Hines, Initial Determination 2025-07-10.
• Commission final determination: 2026-02-26. Per Insta360's public disclosures, the Commission found the asserted claims of the '832 patent invalid. (Insta360 characterized the overall result as prevailing on all six asserted patents; only a design patent remained partially adverse, subject to a redesigned product.)
• URL in the patent record: https://portal.unifiedpatents.com/litigation/International%20Trade%20Commission/case/337-TA-1400

d) District court — GoPro's parallel action in the U.S. District Court for the Central District of California, No. 8:24-cv-00681, stayed 2024-06-03; per Insta360's announcements, the stay was expected to lift after the ITC case concluded.

5. Explicit uncertainties and caveats

1. Date inconsistency in the record. Your header states the current date is April 26, 2026, but the CAFC docket entry for 26-1859 shows a filing date of May 21, 2026, and my operating environment date is 2026-09-30. I cannot reconcile these; the CAFC filing date and any post-filing activity should be verified directly against PACER/CAFC docket 26-1859.
2. FWD claim-by-claim breakdown is second-hand. I retrieved the FWD outcome and status from PTAB case-law databases and snippets of the FWD and Patent Owner Response, not the full FWD PDF from PTAB's own site. The precise allocation (fully unpatentable vs. surviving claims) and the specific grounds should be confirmed against the official Final Written Decision, Paper in IPR2024-01052.
3. Claim 11 text not retrieved. My description of claim 11 is inferred parallelism, not a verified quotation. Do not rely on it as claim language.
4. "Active" legal status ≠ validity. The Google Patents status reflects administrative/fee status. Both the ITC (final determination, Feb. 2026) and the PTAB (FWD, Dec. 2025, on appeal) have made adverse patentability/invalidity findings against the asserted claims of this patent.
5. One secondary source contains an apparent error. An AI-generated case-law summary described the IPR as involving "Insta360's claims 7–8 and 17–18." The '832 patent is GoPro's patent, so that phrasing is wrong. I have not relied on it.
6. This is a technical/analytical summary, not legal advice or a legal conclusion on validity, infringement, or claim scope.

I'll research litigation involving US 11,336,832. Let me run several searches against patent litigation sources.

Let me confirm the specific case numbers and current statuses.

Litigation involving US Patent 11,336,832

Short answer: Yes — US 11,336,832 (GoPro, "Systems and methods for horizon leveling videos") is at the center of a multi-front dispute between GoPro, Inc. and Arashi Vision Inc. d/b/a Insta360 (and its U.S. subsidiary). There are three co-pending/related U.S. proceedings (an ITC § 337 investigation, a stayed district court case, and a PTAB IPR), plus an appellate proceeding.

Note on scope: the searches returned multiple GoPro/Insta360 suits. Several involve different GoPro patents. I have limited the list below to matters in which the '832 patent itself was asserted or challenged. (For example, PTAB IPR2025‑00017 concerned U.S. 10,529,052, and the Chinese actions by Insta360 against GoPro's China entities concern Chinese patents — neither is litigation on '832.)

1. ITC Section 337 Investigation — In the Matter of Certain Cameras, Camera Systems, and Accessories Used Therewith

Interim milestones specific to '832: on Jan. 31, 2025 the Commission declined to review ALJ Initial Determination No. 24, which terminated the investigation as to claims 1‑3, 5‑7, 9 and 10 of the '832 patent. GoPro later dropped/withdrew certain other asserted claims during the investigation.

Sources: https://cacs.mofcom.gov.cn/article/ajycs/ssqdc/[202602](/patent/202602)/[187303](/patent/187303).html ; https://www.mofo.com/resources/news/[260226](/patent/260226)-insta360-section-337 ; https://www.insta360.com/blog/news/insta360-complete-victory-itc-final-ruling.html ; https://law.asia/insta360-gopro-section-337-patent-probe/

2. District Court — GoPro, Inc. v. Arashi Vision Inc. d/b/a Insta360, et al.

Sources: https://www.law.com/radar/card/pm-52908182-gopro-inc-v-arashi-vision-inc-d-b-a-insta360/ ; https://www.ptablitigationblog.com/applying-fintiv-to-a-parallel-itc-investigation/ ; https://ai-lab.exparte.com/case/dct/cacd/8:24-cv-00681/doc/1

3. PTAB Inter Partes Review — Arashi Vision Inc. (d/b/a Insta360) v. GoPro, Inc.

Sources: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-01052 ; https://portal.unifiedpatents.com/ptab/case/IPR2024-01052

4. Federal Circuit appeal

The Google Patents "litigation" record for US 11,336,832 lists a U.S. Court of Appeals for the Federal Circuit proceeding, Appeal No. 26‑1859, which on its face corresponds to the post‑2026 appeal(s) from the ITC final determination and/or the PTAB IPR2024‑01052 final written decision.

Source: https://portal.unifiedpatents.com/litigation/Court%20of%20Appeals%20for%20the%20Federal%20Circuit/case/26-1859

I was unable to independently verify the parties, docketing date, or the specific order being appealed in 26‑1859, so treat those specifics as unconfirmed.

Summary table (US 11,336,832 only)

Caveats and open items

• No other US litigation on the '832 patent surfaced beyond the GoPro–Insta360 family above. I found no case in which '832 was asserted by or against a party other than GoPro/Insta360.
• The Chinese actions (Insta360's suits against GoPro entities in the Jiangsu High Court, filed July 2024, ~RMB 110.2M; and in the Shenzhen Intermediate Court, filed Nov. 2024, ~RMB 40.2M) are not litigation on US 11,336,832 — they assert Insta360's Chinese patents — but they are part of the same commercial dispute and are noted for context.
• A few third-party/AI‑aggregated pages (e.g., Arctic Invent) describe the IPR as involving "Insta360's claims 7‑8 and 17‑18." That characterization is suspect: the '832 patent is GoPro's patent, and the ITC claim‑termination list for '832 runs only to claim 10. I have therefore not repeated those claim numbers as fact.
• The ITC/litigation‑portal links in the Google Patents record (Unified Patents and Darts‑IP "litigation" entries) are aggregator metadata; the underlying dockets are the authorities cited above.

If you need a certified docket check, the ITC EDIS docket for 337‑TA‑1400, the C.D. Cal. docket for 8:24‑cv‑00681, and PTAB Patent Trial & Appeal Case Tracking System for IPR2024‑01052 (plus the Fed. Cir. docket/CM‑ECF for 26‑1859) are the primary sources to pull.

I'll start by researching the PTAB proceeding referenced in the patent record (IPR2024-01052) and any others.

Let me dig deeper into the FWD, institution decision, and appeal.

PTAB Proceedings — U.S. Patent No. 11,336,832 ("Systems and methods for horizon leveling videos")

Sourcing note (read first). The structured "PTAB proceedings on file" block supplied with this task reports that the USPTO Open Data Portal returns no AIA trial proceedings for this patent as of the most recent ingest. That ODP result is stale/incomplete. Web search surfaces a live, fully tried IPR on this exact patent — IPR2024-01052 — and Google Patents' own record for US11336832 flags "PTAB case IPR2024-01052 filed (Final Written Decision)" under Family has litigation. I am therefore reporting one proceeding, flagged as not present in the ODP feed, and I am distinguishing throughout between what I verified from primary/quasi-primary sources and what comes from third-party aggregators.

Proceedings overview

**One AIA trial proceeding on file: IPR2024-01052 (Arashi Vision / Insta360 v. GoPro), status "Final Written Decision – Appealed" — the Board held the challenged dependent claims 7, 8, 17, and 18 unpatentable, but the proceeding is not yet final: GoPro sought Director Review and Insta360 has appealed to the Federal Circuit (26-1859), so no claims have been canceled by certificate yet. Bottom line for a defendant: the patent is wounded, not dead. The Board's § 103 roadmap (Thomason + Pacurariu + Watanabe; Voss + Derbanne) is public and reusable against the size/zoom and location limitations, but the independent claims appear to have survived the trial, so do not build a defense on the assumption that this patent is invalid — and do not tell a court that claims 7/8/17/18 are "canceled," because they are not.

Counts: 1 total — 0 claims-invalidated-as-final, 0 claims-sustained-as-final, 0 settled, 0 institution denied, 1 active-on-appeal. (Breakdown by status per the aggregators' verbatim label: Final Written Decision – Appealed, outcome Unpatentable.)

IPR2024-01052 — Arashi Vision (U.S.) LLC (d/b/a Insta360) v. GoPro, Inc.

• Type: Inter Partes Review
• Filed: 2024-06-20 (petition filing date per ipverse and Ex Parte AI Lab)
• Status: "Final Written Decision - Appealed" (verbatim, aggregator records). Plain English: the Board decided the trial, the FWD is on appeal, and the claims adjudicated unpatentable have not been canceled because § 318(b) cancellation only follows exhaustion of appeal.
• Judge panel: William V. Saindon, Nabeel U. Khan, and Brian D. Range, Administrative Patent Judges (confirmed by the Board's per curiam order captioned across IPR2024-00984/01052/01164/01165, Docket Alarm copy). Counsel: Alex Yap et al. (Morrison & Foerster) for Insta360; James Glass / Patrick Schmidt et al. (Quinn Emanuel) for GoPro.
• Petition grounds: § 103 obviousness, multiple grounds, including:
  • Ground 1 — Thomason in view of Pacurariu (Petition, Paper 2, at 8; Ex. 1004 Thomason, Image Rotation Correction for Video or Photographic Equipment; Ex. 1002 Katsaggelos Declaration). This ground targeted the core horizon-leveling/viewing-window claims.
  • Thomason + Pacurariu + Watanabe directed to the distortion limitations of claims 3–5 and 13–15 (Petitioner's Opposition to Patent Owner's Motion to Strike, asserting Watanabe Ex. 1006 Figs. 5/51/54).
  • Voss in view of Derbanne, directed to the "location of the viewing window … determines framing" limitations of claims 10 and 20, relying on Voss's "inset rectangle" / pillar-boxing disclosure.
  • Claim scope in play ran to at least claims 1 (independent), 3–5/13–15, 7/17, 8/18, and 10/20. I could not retrieve the full FWD text, so I am not representing that any claim outside that list was or was not decided.
• Institution decision: Instituted 2025-01-07 (aggregator records; ~6.5 months after filing). The Decision on Institution (Paper 10) is cited in the record at pages 19–20 for the Board's refusal to import a specification embodiment into claim 1: "Patent Owner does not, however, explain why claim 1's scope must be limited to this embodiment." I did not locate a public partial-denial or denial of any ground at institution; treat "instituted as to the challenged grounds" as reported, not verified line-by-line.
• Final Written Decision: 2025-12-02 (Board's FWD and termination date per aggregator records; FWD issued roughly five weeks ahead of the statutory one-year bar measured from the 2025-01-07 institution). Verdict as reported: the Board held dependent claims 7, 8, 17, and 18 unpatentable (the "size of the viewing window" / "size changes … to simulate changes in zoom" limitations), and the overall proceeding outcome is recorded as "Unpatentable." The Board's reasoning, as characterized in the parties' briefing that preceded the FWD, turned on Thomason's frame-to-frame cropping and re-sizing teaching a viewing-window size change that renders the tilted-camera scene "more zoomed-in" than the level-camera scene — i.e., a simulated zoom (Petition/Reply mapping Thomason FIG. 3d–3e). Caveat: I could not verify the complete claim-by-claim disposition of the FWD (e.g., the Board's treatment of independent claim 1 and of the distortion and location claims). Secondary sources indicate GoPro challenged only the claims 7/8/17/18 findings on Director Review, which implies the other claims did not go against GoPro — but that is an inference, not a verified quotation from the FWD, and it should be confirmed against the FWD (Paper ~44) on PTAB E2E/PTACTS.
• Director Review: GoPro filed a Request for Director Review asking the Director to reverse the unpatentability findings on claims 7, 8, 17, and 18, arguing (i) the Board treated Thomason as a mere suggestion of frame-by-frame window sizing without a proper motivation-to-combine analysis, (ii) the Board disregarded unrebutted evidence that dynamic sizing produces a "pumping" artifact (an unpredictable-effect / teaching-away theory), and (iii) the Board undervalued GoPro's commercial-success evidence tied to its Horizon Leveling feature and misconstrued "viewing window" (source: Arctic Invent case note — third-party, AI-generated summary; disposition of the Director Review request is unknown to me).
• Settlement / termination: No settlement. The record shows a fully litigated trial through sur-reply, motions to strike, and cross-examination (e.g., Katsaggelos depositions taken 2025-03-28 and 2025-04-11, and joint motions to seal GoPro business/financial exhibits, granted 2025-01-06). Several GoPro exhibits (Ex. 2008, 2018, 2019) remain under seal as confidential business information.
• Appeal: Yes — CAFC No. 26-1859, Arashi Vision US LLC v. GoPro Inc, filed 2026-05-21, appeal from PTAB originating in IPR2024-01052; outcome pending (Ex Parte AI Lab CAFC docket). Note the appellant is the petitioner (Insta360), not the patent owner — consistent with a mixed FWD in which Insta360 lost on claims it wanted invalidated and GoPro lost on claims 7/8/17/18. ⚠️ Timeline discrepancy to verify: a 2025-12-02 FWD would ordinarily make the Fed. Cir. notice of appeal due ~2026-02-03 (Fed. Cir. R. 4 / 28 U.S.C. § 2107), yet the docket shows filing on 2026-05-21. The appeal may run from a later decision (e.g., a Director Review or rehearing ruling), or one of the secondary dates is wrong. Confirm on PACER and PTAB E2E before relying on this for any deadline computation.
• Defensive value: The FWD hands a defendant a public, court-tested obviousness theory — Thomason's frame-by-frame rotation correction and re-cropping, combined with Pacurariu's smoothing and Watanabe's distortion teaching — that the Board accepted against the viewing-window size and simulated-zoom limitations, plus a Voss/Derbanne theory against the "location … determines framing" limitations. But because the independent claims apparently survived and the case is on appeal, this patent is not a paper tiger: an IPR-only defense aimed at the independent claims is materially harder than the 7/8/17/18 result suggests.

Strategic summary

Claim status (as best the public record supports). Using the '832 patent's 20-claim structure — independent claim 1 and a second parallel claim set through claim 20, with mirrored limitations at 3–5/13–15 (distortion), 7/17 (determining a size), 8/18 (size changes simulate zoom), and 10/20 (location determines framing):

There is no second IPR on the '832 patent. Do not confuse this docket with the same petitioner's parallel IPRs on different GoPro patents: IPR2024-00984 (U.S. 10,015,413), IPR2024-01164 and IPR2024-01165 (U.S. 10,958,840) — those do not affect the '832 claims but are relevant to a portfolio-wide defense.

Estoppel landscape. § 315(e)(2) estoppel runs against Insta360 and its privies only. If you are a different accused infringer, you are not barred, and the following grounds remain fully available: Thomason (Ex. 1004) alone or with Pacurariu, the Thomason–Pacurariu–Watanabe combination against the distortion claims, and Voss + Derbanne against the location claims — all of which are now public and battle-tested. Conversely, if you are in privity with Insta360 (e.g., a contract manufacturer, distributor, or customer supplied by Insta360), estoppel is a live risk and you should assume those grounds are gone in the district court and in the ITC. Note also that the Board declined to strike Insta360's Reply theories (the motion-to-strike skirmish over Thomason/Pacurariu/Watanabe and "obvious to try," resolved by Board email authorizations on 2025-08-19 and 2025-09-16), so those ground formulations are record-tested rather than new.

Pattern signals. (1) Same-petitioner serial attacks: Insta360 filed four IPRs against GoPro in mid-2024 spanning three patents; this is coordinated IPR-as-litigation-defense, not a defensive aggregator. Unified Patents appears only as a docket-data aggregator in the patent record's litigation links — it is not a petitioner here; there is no aggregator-funded IPR on this patent. (2) Patent owner posture: GoPro litigates hard on both sides — it commissioned the Director Review request and, one presumes, will defend the CAFC appeal, while simultaneously pressing the ITC and C.D. Cal. actions. (3) Parallel-track outcome: in ITC Inv. No. 337-TA-1400, the Commission (vote 2026-02-26, notice published 2026-03-03) declined review of the Final ID's finding of no violation as to the '832 patent and terminated the investigation with an exclusion order only on the design patent D789,435 — so the '832 patent has now lost or been neutralized on multiple fronts. Relevant, but it is an ITC determination, not a PTAB one, and it does not cancel claims.

Family/continuation risk. Cancellation of '832 claims would not extract GoPro from the space: the '832 family continues through U.S. 12,302,000 B2 (from application 17/744,949) and U.S. Pub. 2024/0348927 A1 (from application 18/750,643, filed 2024-06-21). Assume GoPro can re-assert horizon-leveling subject matter through progeny claims.

Recommended next steps

1. Pull the actual FWD (Paper ~44, 2025-12-02) before relying on anything above. Go to PTAB E2E (https://ptab.uspto.gov) → IPR2024-01052, and the PTACTS public docket (https://ptacts.uspto.gov). Confirm the exact claim-by-claim disposition — specifically whether independent claim 1 (and claim 11/12, if present as an independent claim) was held unpatentable or sustained, and whether any of claims 2–6, 9–16, 19–20 were reached. The entire defensive posture flips on that one fact.
2. Do not plead that claims 7, 8, 17, and 18 are "canceled." They are held unpatentable subject to appeal; no § 318(b) certificate has issued. Pending the CAFC appeal (26-1859), they remain enforceable claims, and an assertion built on them is not "sanction bait" — yet. Instead, cite the FWD's unpatentability findings as persuasive authority in a § 282 invalidity case and as a basis for seeking a stay of any parallel litigation pending CAFC resolution.
3. If you are a defendant being asserted on claims 7/8/17/18, or on claims 10/20 (framing/location), file for IPR immediately. The petition has already been written for you by Insta360; the Board has already credited the Thomason/Pacurariu (and Voss/Derbanne) theories. Watch the 1-year § 315(b) bar from service of the complaint. Unlike the Patent Owner in IPR2024-01052, a new petitioner can also fix the vulnerability the FWD exposed by adding a stronger motivation-to-combine and secondary-considerations attack.
4. If you are a defendant on the independent claims, plan a non-IPR defense. Secondary reporting suggests claim 1 survived, and the parallel ITC record shows the Commission found no violation as to '832 — leverage non-infringement (the ITC's claim-construction-driven findings and any estoppel from the Commission's reasoning), invalidity theories beyond the Thomason family, and the Board's own claim-construction statements in the 2025-01-07 Institution Decision at 19–20.
5. Monitor three milestones. (a) PTAB Director Review decision on GoPro's request as to claims 7/8/17/18; (b) CAFC 26-1859 briefing, oral argument, and disposition (also reconcile the FWD-to-appeal timing discrepancy noted above); and (c) any § 318(b) certificate — that is the only event that converts the FWD's findings into actual cancellation.
6. If you are in Insta360's supply chain, assume § 315(e)(2) estoppel and get a privity opinion; if you are unaffiliated, keep the Thomason, Pacurariu, Watanabe, Voss, and Derbanne references as your ready-made § 103 kit.

Confidence flags: The existence, parties, panel, filing date, institution date, FWD date, appeal docket, and grounds of IPR2024-01052 are well corroborated across multiple sources. The precise claim-level disposition of the FWD is NOT verified against the decision text, and the Director Review disposition is unknown. The claim-numbering/parallel-set structure (1↔11) is my inference from the record's paired citations and should be confirmed from the patent's claim set. Treat all aggregator-derived details as needing PTAB E2E confirmation before use in a filing.

I'll research the assignment record and litigation history for US 11336832 before writing up the analysis.

Ownership & Assignment Analysis — US 11,336,832 B1

Patent: Systems and methods for horizon leveling videos · Appl. No. 17/006,536 · Filed 2020-08-28 · Issued 2022-05-17 · Priority 2019-08-30 · Current assignee of record: GoPro, Inc.

Source limitation, stated up front: The USPTO Assignment Center record for this patent could not be pulled in this session — that interface is JavaScript-driven and the reel/frame and correspondent-of-record fields were not exposed in any indexed or cached copy I could reach. Google Patents' legal-events feed records the same five conveyances (assignor, assignee, conveyance type, date) but omits reel/frame and correspondent. I have not invented reel/frame numbers or correspondent names. Where those fields are required below, they are marked not retrieved rather than guessed. Verification path is given under Verdict.

Inventors

All three were GoPro employees at filing. Evidence: the inventors executed an "Assignment of Assignors' Interest" to GoPro, Inc. recorded on the filing date (2020-08-28), with the assignor field listing exactly these three names. This is a standard pre-filing employment assignment — there is no third-party inventor, no university, and no joint-development party in the chain.

Unusual-pattern check — negative. The warning pattern you flag (all inventors departing the original assignee within 12 months of filing, ahead of a fire-sale) is not observable and not indicated here. Two affirmative counter-indications: (a) GoPro continued to prosecute the family internally, adding US 17/744,949 (filed 2022-05-16, later US 12,302,000 B2) and US 18/750,643 (filed 2024-06-21, US 2024/0348927 A1) — a divesting assignee does not typically keep filing continuations on the same disclosure four years out; (b) no inventor was reassigned to a third party at any point; the only post-2020 recorded events are lien filings by GoPro as assignor.

Original assignee

GoPro, Inc. (San Mateo, CA; NASDAQ: GPRO), a/k/a Woodman Labs historically. Named on the face of the issued patent; original and still current owner of record.

• Did they ship a product embodying the claims? Yes, and this is documented in the litigation record rather than inferred. The ITC record for Inv. No. 337-TA-1400 includes the document "HERO9 Black – In-Camera Horizon Leveling," produced as GOPRO_00059907. The '832 patent has a 2019-08-30 priority date and the HERO9 Black (2020) is the first GoPro shipping in-camera horizon leveling implementation. GoPro also asserted this patent as one of six asserted IP rights in its own name. See ITC Initial Determination, 337-TA-1400, July 10, 2025 (ALJ Doris Johnson Hines) and govinfo FR notice 2026-04156.
• Primary line of business: Consumer action cameras, camera systems, accessories, and companion editing software.
• Current status: Operating, but in acute financial distress. On 2025-08-04 GoPro raised a $50M second-lien secured term loan from Farallon Capital Management and granted a first-priority security interest in all IP registrations/applications to Wells Fargo as agent — press release at GoPro IR and Nasdaq. Subsequent 8-K/SEC disclosures contain a going-concern warning, and the company has undertaken layoffs and stated it is seeking financing, buyers, or a merger. Not in bankruptcy and not dissolved — no Chapter 7/11 filing appears in any source, despite sensationalist press framing of the going-concern language.

Assignment timeline

Five recorded events. All reel/frame and correspondent fields: not retrieved — see source limitation above. Dates below are the event dates carried in GoPro/Google Patents legal-events records; where execution and recording dates diverge, the divergence is flagged.

• 2020-08-28 (executed on or before) / recorded 2020-08-28 — Reel not retrieved

  • Conveyance: Assignment (Assignment of Assignors' Interest)
  • Assignor: Daryl Stimm; Kyler William Schwartz; Jonathan Leland Thorn (individuals)
  • Assignee: GoPro, Inc.
  • Correspondent: not retrieved
  • Context: Founding employment assignment — inventors' rights transferred to GoPro contemporaneously with the 2020-08-28 filing of Appl. 17/006,536.

• 2020-10-19 / recorded 2020-10-19 — Reel not retrieved

  • Conveyance: Security Agreement (recorded as "SECURITY INTEREST")
  • Assignor: GoPro, Inc.
  • Assignee: JPMorgan Chase Bank, N.A., as Administrative Agent
  • Correspondent: not retrieved
  • Context: Securitization — GoPro's patents pledged as collateral under its then-existing credit facility. This is a lien, not a transfer of title.

• 2021-01-25 / recorded 2021-01-25 — Reel not retrieved

  • Conveyance: Release (recorded as "RELEASE OF PATENT SECURITY INTEREST")
  • Assignor: JPMorgan Chase Bank, N.A., as Administrative Agent
  • Assignee: GoPro, Inc.
  • Correspondent: not retrieved
  • Context: Lien release — collateral freed, title confirmed back in GoPro. Net effect of the 2020-10-19 and 2021-01-25 pair is zero change in ownership.

• 2025-08-04 / recorded 2025-08-04 — Reel not retrieved

  • Conveyance: Security Agreement (recorded as "SECURITY INTEREST")
  • Assignor: GoPro, Inc.
  • Assignee: Wells Fargo Bank, National Association, as Agent
  • Correspondent: not retrieved
  • Context: Securitization — first-priority security interest in all GoPro IP registrations and applications, granted under the Amended and Restated Guaranty and Security Agreement entered in connection with the Farallon facility.

• 2025-08-04 / recorded 2025-08-04 — Reel not retrieved

  • Conveyance: Security Agreement (recorded as "SECURITY INTEREST")
  • Assignor: GoPro, Inc.
  • Assignee: Farallon Capital Management, L.L.C., as Agent
  • Correspondent: not retrieved
  • Context: Securitization — second-lien security interest in substantially all GoPro assets including IP, securing the $50M term loan. Same lender-side transaction as the Wells Fargo entry above; the two were executed the same day and are intercreditor-linked, not independent transfers.

No post-issuance assignment of title exists in this chain. Every recorded event after the 2020 inventors' assignment is either a lien or the release of a lien. GoPro has held bare title since 2020-08-28.

Timeline diagram

timeline
    title Ownership of US 11336832
    2019 : Priority date Aug 30
    2020 : Application filed Aug 28
         : Inventors assign rights to GoPro
         : JPMorgan records security interest Oct 19
    2021 : JPMorgan security interest released Jan 25
    2022 : Patent issued May 17
    2024 : GoPro sues Insta360 in California
         : ITC investigation 337-TA-1400 opened
         : Insta360 files IPR2024-01052
    2025 : Wells Fargo and Farallon security interests
         : ITC initial determination Jul 10
    2026 : ITC final determination Feb 26

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
No assignment of title has ever run to an entity with an "IP / Patents / Licensing / Holdings / Ventures" suffix. The four post-founding conveyances name only GoPro (as assignor for liens), JPMorgan Chase (2020), Wells Fargo (2025), and Farallon (2025) — all of which are secured lenders taking collateral, not purchasers of the patent. A security interest under UCC Article 9 does not convey ownership; GoPro remains the assignee of record throughout. There is no anonymous LLC anywhere in this chain, so the "no products / registered-agent address / single-purpose LLC" test never gets triggered.

2. Known asserter in the chain — NOT PRESENT.
The current and sole owner is GoPro, Inc., a publicly traded operating company. GoPro does not appear on the RPX or Unified Patents high-frequency NPE plaintiff directories, and Unified Patents' involvement here is merely as the data publisher for the PTAB docket (Unified Patents portal, IPR2024-01052) — not as an asserter or owner. None of Acacia, Marathon, IV, IPNav, Wi-LAN, Conversant, Vringo, Pendrell, Round Rock, etc. appears anywhere in the chain.

3. Repeat correspondent across the chain — UNRESOLVED / not retrieved.
This is the one signal I cannot responsibly score. The correspondents of record on the 2020 JPMorgan filing, the 2021 release, and the two 2025 lender filings were not exposed by any source I could reach. I am flagging this as a genuine open item rather than calling it absent, because it is precisely the field your brief identifies as the highest-value tell. Practical note for follow-up: in operating-company securitization chains the correspondent is typically the lender's counsel (a large general-practice firm) rather than the patent owner's prosecution firm, which would make recurrence across the 2020 and 2025 entries an intercreditor-administration artifact and not an NPE tell. That is a hypothesis to test against the reel records, not a finding — and, importantly, the recurrence test you describe (one lawyer shepherding a family of unrelated shell LLCs) cannot be met on this chain anyway because there are no shells.

4. Cascading transfers — NOT PRESENT.
Zero chained LLC transfers. The only clustered dates are 2025-08-04 (Wells Fargo + Farallon), which are the first- and second-lien halves of a single intercreditor-arranged financing, executed the same day against the same assignor for the same credit facility. That is a securitization event, not a cascade. Spacing between the only two title-relevant events (2020 assignment, then nothing) is measured in years, not months.

5. Pre-litigation transfer — NOT PRESENT.
GoPro filed the ITC complaint on 2024-03-29 and the parallel California Central District action (8:24-cv-00681) in 2024. The last title-transfer-adjacent record before suit was the 2021-01-25 lien release, more than three years earlier. There is no assignment within six months of the first suit — and, decisively, no assignment to the plaintiff that needed to be arranged at all, because GoPro had owned the patent outright since 2020. The clean-standing/venue-priming rationale for a pre-suit transfer is simply absent.

6. Bankruptcy fire-sale — NOT PRESENT (but distress risk is elevated).
No Chapter 7 or Chapter 11 filing by GoPro appears in any source, and no patent sale in a bankruptcy proceeding has occurred. Elevated going-concern language in GoPro's 2025–2026 SEC filings plus repeated layoffs mean a future §363 sale is a live risk, not an event. Should that change, this patent (and the rest of GoPro's IP) would be a candidate asset sale — worth re-running this analysis if a Chapter 11 appears on the docket.

7. Privateering — NOT PRESENT.
GoPro asserted in its own name, for its own commercial benefit, against a direct product competitor. GoPro's ITC complaint and parallel district court suit name Arashi Vision Inc. (d/b/a Insta360) and Arashi Vision (U.S.) LLC as respondents/defendants. There is no NPE intermediary asserting on GoPro's behalf, no revenue-share arrangement surfaced in SEC filings, and no Patent Progress/EFF coverage of a privateering arrangement — because there is nothing to cover.

8. Defensive aggregator — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at an operating company that actively litigated.

Verdict

Operating-company assertion

GoPro, Inc. has held unbroken title to US 11,336,832 since the inventors' assignment recorded 2020-08-28, and every subsequent recorded conveyance — JPMorgan 2020-10-19, its release 2021-01-25, and the twin Wells Fargo / Farallon security interests on 2025-08-04 — is a lien or lien-release naming GoPro as assignor, not a transfer of ownership. GoPro then asserted the patent in its own name in ITC Inv. No. 337-TA-1400 and parallel C.D. Cal. 8:24-cv-00681, against a genuine product competitor (Insta360), while shipping a commercial product (HERO9 Black in-camera horizon leveling) that practices the claims — the textbook definition of operating-company assertion, and the inverse of every NPE signature in this framework.

Attrition note for completeness (this is the litigation outcome, not an ownership event): the assertion did not succeed. The ALJ's July 10, 2025 initial determination found no violation as to the '832 patent; the Commission declined review on September 11, 2025, making no-violation final; and Insta360's IPR2024-01052 produced invalidation of the '832 claims, which GoPro took to Director Review on 2026-01-05. Insta360's February 2026 announcement describes the ITC final determination as clearing it on all five utility patents including the '832 (Insta360; govinfo 2026-04156). GoPro's right of appeal ran 60 days from the final determination, consistent with the CAFC docket 26-1859 appearing in the family's litigation record. None of this changes the ownership conclusion — but it means this patent is a depreciated asset in a distressed balance sheet, and the only scenario in which this chain becomes interesting for NPE purposes is a future GoPro Chapter 11 §363 sale.

Verify at: USPTO Assignment Center → search property 11336832; cross-check the reel/frame and correspondent fields against Google Patents legal events for US11336832B1. Those two fields are the outstanding gap in this report.

I'll research US Patent 11336832 and its prior art citations.

Let me get more detail on the cited references and the IPR grounds.

Let me identify the full citations for the IPR prior-art references and confirm the patent's citation list.

Scope and sourcing note

Per your instruction, I searched specifically for US 11,336,832 (no substitutions). The patent identifies literally as US 11,336,832 B1 (also written US11336832B1). One source I retrieved (a district-court complaint) misstates the issue date as "May 17, 2020"; the authoritative record (Google Patents, USPTO data) shows grant/publication date 2022-05-17. I flag the discrepancy rather than silently correcting it.

Two distinct prior-art populations exist for this patent, and they are different:

• (A) The references cited on the face of the patent (the "References Cited"/Citations list).
• (B) The references asserted against it in litigation (IPR2024-01052, ITC 337-TA-1400) — these were not the examiner's citations.

I treat both, and I distinguish my own § 102 analysis from official findings.

1. Patent under analysis

Because the effective priority date is 2019-08-30, any reference published or filed before that date is available under 35 U.S.C. § 102(a)(1)/(a)(2).

2. GROUP A — References cited on the face of the patent

Using the Google Patents citation data retrievable for US11336832B1, the cited references are three:

A-1. DE 10003732 A1 — Robert Bosch GmbH

• Title/Subject: Mobile video communication device for editing image data; detects a rotational angle difference between the directed image axis and a preferred direction, and image data is corrected based thereon.
• Publication date: 2001-08-16 (well before the 2019 priority date).
• Brief description: Discloses correcting captured image data in a mobile video device using a measured rotational-angle offset — i.e., sensor-derived orientation used to rotate/correct imagery. It does not, on its face, describe a viewing-window/punchout that is redetermined per-frame as a function of progress.
• Potential § 102 relevance: Potentially anticipatory only of the broadest "obtain image data + obtain rotational information + correct based on rotation" concepts, i.e., it is a candidate against claim 1/11 only if those claims were read to omit the per-progress-length viewing-window determination — which the patent owner's IPR briefing expressly disputes. Realistically a § 102(a)(2) art reference for the rotation-correction element, not for the punchout limitation. It is stronger as § 103 background.

A-2. US 2010/0245601 A1 — Casio Computer Co., Ltd.

• Title/Subject: Image recording apparatus, image tilt correction method, and recording medium storing image tilt correction program.
• Publication date: 2010-09-30.
• Brief description: Discloses detecting image tilt and correcting the recorded image's tilt — i.e., horizon/tilt correction of imagery. Appears directed to tilt correction generally, and does not appear to disclose a dynamically sized, per-frame punchout window.
• Potential § 102 relevance: Candidate against the tilt/rotation-correction elements underlying claim 9/19 (rotation of the viewing window) and the general leveling element of claim 1/11, if construed broadly. Like A-1, it does not appear to disclose the "as a function of progress through the progress length" viewing-window determination that the patent owner treats as the point of novelty.

A-3. US 2012/0105572 A1 — Sammon, Russell P.

• Title/Subject: Automatically adjusting a video-capture device.
• Publication date: 2012-05-03 (filed 2010-10-28).
• Brief description: A video-capture device (pocket camcorder/smartphone) automatically changes settings when placed on a supporting surface, enabling features such as auto-framing, subject tracking, and digital tilt adjustment; determines device state from sensors and/or from processing captured video to quantify shake.
• Potential § 102 relevance: The "digital tilt adjustment" and auto-framing disclosures are the closest of the three to the placement/framing concepts in claim 10/20 (location of the viewing window determining framing) and possibly the location element of claim 1/11. However, Sammon's framing is not driven by per-frame rotational-position information across a capture duration, so a clean § 102 anticipation of claims 1/11 or 10/20 is not apparent from the disclosure as described.

Caveat on Group A completeness: The retrievable Google Patents citation table for this family lists three references (the three above). I was not able to fully verify whether the printed USPTO "References Cited" section of the '832 patent contains additional applicant-cited or examiner-cited entries beyond these three. Treat Group A as "the citation list as reflected in the retrieved Google Patents data," not as a certified face-of-patent list.

3. GROUP B — Prior art asserted against the '832 patent in IPR2024-01052 / ITC 337-TA-1400

The Petition (Paper 2) framed four grounds, all under § 103, with the following claim coverage:

B-1. Thomason — "Image Rotation Correction for Video or Photographic Equipment"

• Full citation (as verified): EP 1 500 045 A1 (published 2005-01-26); granted as EP 1 500 045 B1; inventor Graham G. Thomason; applicant/owner Koninklijke Philips Electronics N.V.; GB priority 2002-04-15; DE family member DE 603 09 014 T2 (2007-05-02). Australian counterpart app. 2003214553, filed 2003-04-03.
• Brief description: A camera (2) with a lens (4) and CCD (6) and a rotation sensor (14) — an accurate accelerometer measuring camera angle to the vertical (accuracy of 5° or better) — stores image data (8) together with the measured angle, and on replay a processor (22) applies a rotation correction so that "true horizontals appear as horizontal," then crops the rotated image to eliminate blank space (FIGS. 3c–3d, frames 40/42), resizing so the final frame equals the original frame size (FIG. 3e). It describes capturing a "sequence of items of image data 8 representing sequential frames of the moving image," and contemplates circular sensor arrays and 1:1 frame aspect ratios.
• § 102 / § 103 relevance — this is the most dangerous reference:
  • The ITC ALJ found, by clear and convincing evidence, that claim 1 is anticipated by Thomason, and that claims 7 and 8 are anticipated by Thomason, and that claims 3 and 4 are obvious over Thomason + Watanabe (Commission declined review of the invalidity ID).
  • So under § 102: claim 1, claim 7, and claim 8 are the specific claims asserted/substantiated as anticipated by Thomason.
  • The patent owner's contrary positions: Thomason discloses only a one-time orientation measurement and correction (relying on singular "rotation angle" in ¶[0032]), and no express or inherent frame-by-frame horizon leveling; and its single accelerometer "can, at best, crudely infer rotation about just a single axis." The PTAB nevertheless held claims 7, 8, 17, 18 unpatentable as obvious (Final Written Decision, 2025-12-02), which GoPro has challenged by Request for Director Review and on appeal.

B-2. Pacurariu — relied on in Grounds 1 and 2, paired with Thomason for claims 1, 2, 7–12, 17–20. The IPR record treats it as supplying viewing-window/punchout aspects (contrasted in GoPro's POR as failing to disclose a viewing window that is dynamic in size, rotation, and location, claims 7–8/17–18, 9/19, 10/20).

• I was not able to verify Pacurariu's full citation (patent/publication number, date, assignee, title) from the sources retrieved in this session. I will not supply a number I cannot confirm.

B-3. Voss — relied on in Grounds 3 and 4 with Derbanne. The record discusses Voss's "inset rectangle," its cropping vs. "letter boxing"/"pillar boxing" disclosures (e.g., Voss at 10:20–21, 11:5–7, and the Abstract), and an "oscillat[ing]" example.

• Full citation not verified in this session (number/date/assignee unconfirmed). Do not rely on an unverified number.

B-4. Derbanne — relied on in Grounds 3 and 4 with Voss.

• Full citation not verified in this session.

B-5. Watanabe — relied on only in Grounds 2 and 4, against the distortion claims 3–6 and 13–16. GoPro characterizes it as disclosing removal of distortion before cropping and adding a different distortion based on the new crop — i.e., not preserving the input distortion through horizon leveling, and not involving horizon leveling at all. This is the reference whose sufficiency is most contested (GoPro moved to strike Petitioner's attempt to prove claims 3–5/13–15 with Thomason–Pacurariu without Watanabe as an improper new ground).

• Full citation not verified in this session.

4. Bottom line — the most relevant prior art, ranked

1. Thomason, EP 1 500 045 A1 (pub. 2005-01-26; GB priority 2002-04-15) — by far the most relevant. It is the only reference asserted with an actual § 102 anticipation finding against specific claims of the '832 patent: claim 1, claim 7, and claim 8 (ITC ALJ ID, affirmed on non-review); and per the PTAB FWD, claims 7, 8, 17, 18 were held unpatentable. Its sensor-derived rotation measurement + rotation correction + cropping to a punchout maps onto the '832 patent's core independent-claim elements.
2. Voss + Derbanne (Ground 3–4) — the competing § 103 combination for claims 1, 2, 7–12, 17–20, keyed on Voss's "inset rectangle"/pillar-boxing disclosure.
3. Pacurariu — § 103 partner to Thomason for the same claim set.
4. Watanabe — only relevant to the distortion claims 3–6 / 13–16, and its relevance is disputed.
5. Face-of-patent citations — DE 10003732 A1 (2001-08-16), US 2010/0245601 A1 (2010-09-30), US 2012/0105572 A1 (2012-05-03): none appear to be clean § 102 anticipatory art against independent claims 1/11 as issued, because none is described as redetermining a viewing window as a function of progress through the progress length based on per-frame rotational position. They are better characterized as § 103 background art, with US 2012/0105572 A1 (Sammon) the most relevant of the three to the framing/location concepts of claim 10/20.

5. Explicit confidence and gaps

• High confidence: the '832 patent's bibliographic data, claim structure, the four IPR grounds and their claim/reference mapping, and the ITC finding that claims 1, 7, 8 are anticipated by Thomason.
• High confidence: the Thomason citation (EP 1 500 045 A1 / B1; DE 603 09 014 T2; GB priority 2002-04-15; Philips; inventor Graham G. Thomason).
• Low confidence / unverified: the full citations for Pacurariu, Voss, Derbanne, and Watanabe (numbers, dates, assignees). I deliberately did not invent them.
• Unverified: whether the printed USPTO "References Cited" list on the '832 patent contains entries beyond the three shown in the Google Patents citation table.
• Not asserted: none of the § 102 attributions in Group A are official; they are my own element-by-element reading. The only official § 102 findings are the ITC's as to Thomason.

I'll research the prior art and litigation record for this patent before analyzing.

Let me identify the specific prior art references and the claim set.

Now let me pin down the specific reference numbers and the claim set.

Obviousness Analysis — U.S. Patent No. 11,336,832 B1 ("Systems and methods for horizon leveling videos")

0. A note on the "Prior Art" section of the page

The Google Patents page you supplied does not contain a list of prior-art references. Its "Prior Art" content is limited to:

• Prior art date: 2019-08-30
• Prior art keywords: visual content · viewing window · horizon · progress · video

Because the page itself provides no reference list, I have used the patent's own litigation record that the same page links to — the PTAB proceeding IPR2024-01052 (Petitioner: Arashi Vision Inc. d/b/a Insta360; Patent Owner: GoPro, Inc.), filed 2024-06-20, instituted 2025-01-07, Final Written Decision terminated 2025-12-02, status "Final Written Decision – Appealed" (portal.unifiedpatents.com/ptab/case/IPR2024-01052; ipverse.greyb.com). That record is the authoritative source for which references were actually applied against this patent.

Where I could not verify an identifier, I say so rather than guessing.

1. Temporal framework

All references relied on below published before 2019-08-30 and are therefore §102(a)(1)/§102(a)(2) prior art. Petitioner also represented that "None of the references presented in this Petition — Thomason, Pacurariu, Watanabe, Voss, or Derbanne — were considered during prosecution of the '832 patent," so there is no §325(d) or file-wrapper obstacle to their use.

2. The references (as identified in the IPR record)

3. The claims at issue

The claim listing was not included in the page text, but the Board's FWD and the Petition excerpts establish the following (reconstructed from the record, not from a claim listing I retrieved):

• Claims 1–10 — system claims ("one or more physical processors configured by machine-readable instructions to…"). Claim 1 (at 28:27–65) requires: obtain video info (capture duration, progress length, visual content viewable as a function of progress, field of view) → obtain rotational position information characterizing rotational positions as a function of progress through the capture duration → determine a viewing window as a function of progress through the progress length based on those rotational positions, the window defining extents to be included in horizon-leveled visual content, including determination of placement → generate horizon-leveled visual content including "a punchout of the extents of the visual content defined by the viewing window," inclusion of which "effectuates horizon leveling."
• Claims 11–20 — method analogues (e.g., "the method of claim 11").
• Dependent mapping: 2/12 = one-by-one aspect ratio reducing off‑axis‑horizon impact; 3–6 and 13–16 = distortion (barrel/pincushion; wide‑FOV optical element; distortion applied to non‑distorted content); 7/17 = size determination as a function of progress; 8/18 = size changes to simulate zoom; 9/19 = rotation of the viewing window; 10/20 = location of the viewing window determining framing.

Two claim-construction holdings are outcome-determinative:

1. "Including" ≠ "equate." The Board rejected GoPro's argument that "the 'extents' … must equate to the horizon-leveled visual content." Since X can include Y without equaling it, references that sample a rectangle and then re-project/crop the sampled content (Voss's "output rectangle") still meet the punchout limitation.
2. Capability suffices for "configured to" claims. "The Voss-Derbanne combination's instructions only being executed in a certain mode does not affect whether or not the instructions meet the claim recitation."

4. Grounds of obviousness

Ground 1 — Thomason + Pacurariu (asserted for claims 1, 2, 7–12, 17–20)

• Motivation / rational underpinning (KSR (C),(D)): Both references are in the same field — rotating visual content in video data to compensate for image-capture-device motion — and both target the same problem, amateur mis-orientation ("Cameras … are frequently used by amateurs who may have difficulties in correctly orienting the camera"; "There is thus a need for a camera capable of reducing the effect of poor orientation"). Thomason supplies sensor-derived angle data stored per frame; Pacurariu teaches storing the tilt value as metadata with each frame and determining a viewing window for sequential frames. The combination yields the predictable result that "rotational position information, such as a tilt value as taught by Pacurariu, may be obtained for each video frame in Thomason's video content, be saved as metadata or as part of each video frame …, and be used to generate a rotated image for each video frame" — i.e., effective horizon-leveled video.
• Obvious to try (KSR (E)): Petr. argued a finite number of identified, predictable solutions for storing video-plus-rotation data (series of images vs. compressed, with metadata in various formats).
• Weakness identified in the record: GoPro argued (and the panel explored at the hearing) that Thomason's single-accelerometer, FIG. 3 example is a one-time correction for a video rather than a frame-by-frame, progress-dependent determination, and that Thomason does not support a change of viewing-window size as required by claims 7/8 (and 17/18). Petitioner's own slide-23 argument for claims 7–8 on the Thomason-Pacurariu combination appears not to have been presented in the Petition — a procedural gap GoPro pressed hard at the oral hearing.

Ground 2 — Voss + Derbanne (the combination the Board ultimately credited)

The Board found "Petitioner has shown by a preponderance of evidence that the combination of Voss and Derbanne teaches or suggests all elements of claims 1 and 11" and, "based on objective indicia of non-obviousness as we discuss below," that claims 1 and 11 (and, per the FWD summary, claims 1, 2, 7–12 and 17–20) are unpatentable.

Motivations a POSITA would have had:

1. Same field, same technique (KSR (C)): Both rotate visual content to compensate for device motion; both derive a crop region from a larger capture region using inertial sensor data. GoPro's own expert conceded the equivalence: "We're compensating translation and rotations. We are doing the same things."
2. Improving a known device in a known way (KSR (D)): Derbanne supplies the data architecture (video content with a progress length; per-moment rotational position information; punch-outs) that lets Voss's inset-rectangle horizon-leveling run "as a function of progress through the progress length," yielding "the predictable result of improving the video management and processing capabilities of Voss."
3. Finite, predictable implementation choices (KSR (E)): A limited number of ways to store video frames and associated rotational data; low effort, reasonable expectation of success; "may be combined … without requiring much effort."
4. Design incentives / market forces (KSR (F)): Consumer demand for stabilized, share-ready video from handheld devices rotating freely during capture — the stated purpose of Voss's application itself.

Element mapping (claim 1 / claim 11):

Dependent claims likewise map: 2/12 — Voss expressly discloses 1:1 as a user-selectable output ratio and a 1:1 sampled rectangle, and a POSITA would understand (as the '832's own FIG. 8 discussion confirms) that a square window makes an off-axis horizon less conspicuous than a wide 16:9 window. 7–8/17–18 — Voss's inset rectangle changes size with rotation (max inscribed rectangle shrinks as tilt grows) and interpolates/zooms between ratios. 9/19 — Voss's inverse rotation of the window. 10/20 — Voss's inscribed rectangle 440 as an off-center overlay determining framing.

Ground 3 — Voss + Derbanne + Watanabe (claims 3–6, 13–16) — rejected

Watanabe does not involve horizon leveling, and it removes the input distortion before cropping, then adds a different distortion to the new crop. The claims require distortion that remains in the horizon-leveled content and that reduces the impact of an off-axis horizon. GoPro also argued Watanabe teaches the opposite of the claim (distortion as something to be eliminated). On this record, the distortion claims were not shown to be obvious, and they survived the FWD.

5. Secondary considerations

The Board expressly weighed objective indicia of non-obviousness (GoPro's POR asserted pioneering status, industry praise, and copying by Petitioner: "[Petitioner] is a recent entrant into this market that has found significant success by copying GoPro's designs and innovations, including the patent at issue"). Those indicia did not carry the day for claims 1, 2, 7–12, 17–20: the Board still held them unpatentable. A secondary-considerations defense is only as strong as its nexus to the claimed subject matter, and here the claims were read broadly onto the prior art at the construction stage.

6. Bottom line

1. Strongest §103 attack: Voss + Derbanne. A Board-endorsed preponderance finding of unpatentability for claims 1, 2, 7–12, 17–20 (all claims other than the distortion claims). The Board found both express disclosure and a rational, reasonable-expectation-of-success motivation to combine.
2. Alternative §103 attack: Thomason + Pacurariu, asserted against the same claim set. It supplies a strong teaching of per-frame orientation measurement + crop-to-level + metadata storage, but its vulnerability is the one-time-correction character of Thomason's FIG. 3 disclosure and the absence of a Petition theory for the changing viewing-window size of claims 7/8 (17/18).
3. Claims 3–6 and 13–16 (distortion) are the surviving claims. They were not shown obvious over Voss + Derbanne + Watanabe. An untested alternative theory would substitute for Watanabe a wide-FOV/barrel-distortion action-camera reference (claims 5/15 expressly recite a "wide field of view optical element"), which the '832's own FIG. 7 discussion of barrel distortion supports — but on the present record that theory is unproven, and a "fisheye-lens" reference alone would not obviously disclose the "reducing impact of an off-axis horizon" function.

7. Status caveats (as of the record retrieved)

• IPR2024-01052's Final Written Decision (terminated 2025-12-02) is on appeal — CAFC case 26-1859 is listed on the page; the practitioner summary classifies the FWD as finding claims 1, 2, 7–12, 17–20 unpatentable under §103 while claims 3–6 and 13–16 survived (cases.arcticinvent.com). Note that the Google Patents page still shows the patent as Active with adjusted expiration 2040-11-30 — legal status on that page is expressly an assumption, not a legal conclusion.
• Parallel ITC investigation 337-TA-1400 ran concurrently (page links it); GoPro's parallel district-court case 8:24-cv-00681 (C.D. Cal.) was stayed at 35 U.S.C. §1659 in June 2024 pending the 337 case.
• Publicly reported ITC final determination of 2026-02-26 is said to include a finding that the '832 claims are invalid, alongside non-infringement/invalidity findings on the other asserted utility patents and a limited exclusion order on one design patent (techi.com; 影石创新 SSE announcement, 2026-02-28). Either side may appeal within 60 days of the final determination.
• Identifiers I could not confirm from retrieved sources and therefore do not assert: the document numbers for Pacurariu and Watanabe, and the specific family member constituting Exhibit 1008 (Derbanne). Thomason (US 2005/0168583 A1 / WO 03/088147 A1 / EP 1 500 045 B1) and Voss (US 2015/0348325 A1, related to US 9,942,464 B2) were verified.