US 10015413

I'll run targeted searches for this specific patent number and its appellate docket activity.

Now let me look for the Federal Circuit 2026 appellate dockets referenced in the record.

US 10,015,413 B1 — Verified Identification

I searched for the literal number 10015413 and found no similar-number substitutions needed. Everything below is anchored to the authoritative full text supplied for US10015413B1 (https://patents.google.com/patent/US10015413/en), cross-checked against PTAB, ITC and CAFC docket material from search.

⚠️ One important framing note: this patent's claims are no longer live in the ordinary sense. The Google Patents legal-events record shows statutory disclaimers under 35 U.S.C. § 253 covering claims 2–5, 8–20 and claims 1, 6, and 7 — i.e., collectively all claims 1–20 — with an effective date of 2024-11-05. Independent claims 1, 2 and 13 are therefore disclaimed, not merely invalidated. Details and caveats below.

---

Bibliographic Data

This is a continuation in a long chain: 14/180,887 → US 8,917,329 B1; 14/536,315 → US 9,294,671 B2; 15/011,253 → US 9,674,429 B2; then 15/492,738 → US 10,015,413 B1, with later continuations US 10,368,013; US 10,819,918; US 11,190,709; US 11,722,631; US 12,302,028; and pending US 2025/0267236 A1. Foreign counterparts: EP 3036892 A4, CN 105637855 B, WO 2015/026864 A1.

Record oddity (noted, not corrected): the "Related Parent Applications" table lists US 15/044,253, titled "Method and Apparatus for Indoor Location Estimation Among Peer-To-Peer Devices" — plainly unrelated subject matter, likely a Google Patents data artifact.

---

Abstract (as issued)

A camera system captures an image in a source aspect ratio and applies a transformation to the input image to scale and warp the image to generate an output image having a target aspect ratio different than the source aspect ratio. The output image has the same field of view as the input image, maintains image resolution, and limits distortion to levels that do not substantially affect the viewing experience. In one embodiment, the output image is non-linearly warped relative to the input image such that a distortion in the output image relative to the input image is greater in a corner region of the output image than a center region of the output image.

The specification's problem statement: cropping loses field of view, linear scaling (e.g., 4:3 → 16:9) produces the "short and fat" look, and padding wastes display area. The disclosed fix is a non-linear warp performed in-camera, with distortion pushed toward the frame edges/corners. The description gives explicit polynomial warp equations (a 6th-order Δh/H horizontal function, and a family Δv/V₀ … Δv/V₁₀₀ for vertical warp as a function of horizontal position), plus a lens-distortion relation (FOV) = (21.904)*(IH) − (0.6852)*(IH)^2.

---

Plain-Language Overview of the Independent Claims

Claim 1 — Image capture apparatus (the only independent claim with a display).
A camera apparatus has: (a) an image sensor capturing an input image at a source aspect ratio, with pixels at input positions on two perpendicular axes; (b) a display; and (c) processor(s) programmed to do three things — obtain the input image; determine a portion of the image containing a subset of pixels; and apply a transformation producing an output image at a different target aspect ratio. The transformation must do two specific things: (i) non-uniformly shift pixels from input to output positions based on the input positions along the first axis AND along the second axis (so the shift is a 2-D function of position, not a one-dimensional stretch); and (ii) shift pixels so that pixels inside the "portion" move less than pixels outside it. Finally, the output image is presented on the display.

Claim 2 — System that converts between aspect ratios.
The same subject matter as claim 1, but framed as a system of one or more physical processors — no image sensor and no display required. Same three steps (obtain, determine a portion, apply 2-D non-uniform transformation with the portion moving least) — absent the present-on-display step. This is the broadest apparatus-style claim, and it does not tie the image to in-camera capture.

Claim 13 — Method.
The method counterpart, performed by a computing system with one or more physical processors: obtaining an input image at a source aspect ratio; determining a portion containing a pixel subset; and applying the transformation that yields the different target aspect ratio with the same two limitations (2-D non-uniform shifting; sub-portion pixels shifted less than others). No display, no sensor.

Dependent-claim architecture (maps across claims 2/13 and 1–12):

• Claims 3–5 / 14–15: the transformation decomposed into a uniform scaling along one axis plus a non-uniform warping along the other; optionally the warped/scaled subset's positions are preserved (zero net displacement); scaling is what changes source to target ratio.
• Claims 6–7: processors are inside a camera; output shown on the camera's display.
• Claims 8/16 and 9/17: the "portion" is a center or a non-center region.
• Claims 10/18: shifting is further based on distances between input positions and the portion.
• Claims 11/19: the portion is chosen based on an object's location, so the portion contains the object (dynamic ROI for video).
• Claims 12/20: the input image is a video frame.

Note the claim-1 display requirement, which the PTAB read narrowly in the IPR: claim 1's "media source content" language was construed to cover externally supplied content, not content captured by the device's own camera.

---

Post-Grant Litigation and Validity Status

PTAB — IPR2024-00984, Arashi Vision Inc. (d/b/a Insta360) v. GoPro, Inc. (filed 2024-06-20).
The Board issued a Final Written Decision finding claims 6 and 7 unpatentable as obvious over Meulen (panoramic display for a wireless device) and Chao (digital image auto-resizing), and found claim 1 not unpatentable — distinguishing it because the "media source content" term was read to exclude content captured by the internal camera. Docket and summary sources: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-00984 and https://cases.arcticinvent.com/us/patent/us/judgment-final-written-decision-determining-some-challenged-claims-unp/ipr2024-00984/. Bloomberg Law's automation summary (Dec 1, 2025): the Board found "some challenged claims" obvious over a panoramic-display reference and an auto-resizing reference.

Statutory disclaimer. Google Patents legal events record multiple DC (disclaimer filed) entries with effective date 2024-11-05:

• "DISCLAIM THE FOLLOWING COMPLETE CLAIMS 2-5, 8-20"
• "DISCLAIM THE FOLLOWING COMPLETE CLAIMS 1, 6, AND 7"

Taken together these disclaim every claim, 1 through 20. Uncertainty: the events feed repeats identical descriptions with different posting dates (2025-03-25, 2025-04-08, 2025-05-06, 2025-07-08 for the 2–5/8–20 group; 2025-06-03 and 2025-08-12 for claims 1/6/7) — these look like duplicate propagation of the same filings, so I would not treat each row as a distinct disclaimer. The aggregate effect (all claims disclaimed, effective 2024-11-05) is the substantive point, but the exact number and dates of filings cannot be pinned down from this record alone. Claims 1–20 being disclaimed is also consistent with why the litigation positioned claims 1, 6, 7 as the challenged set and why the FWD's partial outcome became moot commercially.

CAFC 2026 dockets — both live, both from IPR2024-00984:

• 26-1391 — Arashi Vision US LLC v. GoPro Inc, filed 2026-01-28, appeal from PTAB, originating case IPR2024-00984; outcome pending, "Pending Judge Assignment" (https://ai-lab.exparte.com/case/cafc/26-1391/arashi-vision-us-llc-v-gopro-inc).
• 26-1430 — GoPro Inc v. Arashi Vision US LLC, filed 2026-02-12, appeal from PTAB, originating case IPR2024-00984; outcome pending (https://ai-lab.exparte.com/case/cafc/26-1430/gopro-inc-v-arashi-vision-us-llc).

So both sides cross-appealed the same PTAB decision. No merits ruling exists as of the current date.

ITC Inv. No. 337-TA-1400 — Certain Cameras, Camera Systems, and Accessories Used Therewith. GoPro filed the § 337 complaint on 2024-03-29 against Arashi Vision Inc. and Arashi Vision (U.S.) LLC, asserting six IP rights including the '413 patent. A parallel district court action, GoPro, Inc. v. Arashi Vision Inc., No. 8:24-cv-00681-HDE-JCV (C.D. Cal.), was filed the same day and stayed on 2024-06-03 pending the ITC case. Final determination was reported on 2026-02-26: Arashi/Insta360 announced that for the '413 patent the accused products were found non-infringing and the claims invalid or partially invalid, with five of six asserted patents resolved in its favor and only a design patent (D789,435) reaching the accused legacy products. Sources are the issuer's Shanghai Stock Exchange announcements (公告编号 2026-006) and Chinese-language press, e.g. https://www.9fzt.com/detail/sh_688775_9_3e9565ef4d02326f3d706623b546beb9.html and https://news.qq.com/rain/a/20260227A05GIW00. I did not retrieve the primary USITC final-determination document, so treat the ITC outcome as reported by the patent owner rather than independently confirmed.

Also of note in the family: IPR2024-01434 (on design patent D789,435) was not instituted, per the Board's decision reproduced at https://bannerwitcoff.com/wp-content/uploads/2025/04/IPR2024-01434.pdf.

---

Explicit Uncertainty Flags

1. I did not query USPTO PatentCenter or the CAFC PACER system directly — the sources were Google Patents, PTAB-tracking databases, and reported docket summaries. Bibliographic data (filing 2017-04-20, issue 2018-07-03, priority 2013-08-22, inventors, assignee) comes from the supplied authoritative patent text and is reliable. Docket dates come from third-party aggregators and should be re-verified against PACER before being relied on.
2. The list of disclaimed claims is from Google Patents' legal-events feed, which shows duplicated entries; the underlying § 253 filings should be confirmed in the USPTO's official disclaimer records. The practical effect, however — all 20 claims disclaimed — appears clear.
3. The FWD's precise issuance date is not stated in the supplied text; the docket shows Board activity through late November 2025 and Bloomberg Law reported the decision on 2025-12-01. I would state the FWD as "issued late November/early December 2025" rather than giving one date.
4. ITC final determination specifics are per the patent owner's own disclosure, not a primary ITC document.
5. The "anticipated expiration 2034-02-14" figure is as listed by Google Patents; with all claims disclaimed, expiration is largely academic, and I have not independently verified the terminal-disclaimer/PTA arithmetic that produces that date.

If your interest is enforceability rather than technical content: the operative takeaway is that US 10,015,413 has no enforceable claims remaining — claims 1, 6 and 7 were disclaimed after the PTAB upheld claim 1 and struck claims 6–7, and claims 2–5 and 8–20 were disclaimed as well, while the cross-appeals at 26-1391 and 26-1430 remain pending on the PTAB record.

I'll search for litigation involving US 10015413 across patent litigation sources.

Let me research the appellate cases and the district court docket status.

I'll start by reconciling the structured "no PTAB activity" block against what the patent record and the open web actually show.

Now let me pin down the FWD, the disclaimers, and the Federal Circuit appeals.

Let me nail down the FWD specifics, the institution date, and the two Federal Circuit appeals.

PTAB Proceedings — U.S. Patent No. 10,015,413 B1 ("Conversion between aspect ratios in camera")

Threshold reconciliation (read first). The structured "PTAB proceedings on file" block says the USPTO ODP API returns no AIA trial proceedings for this patent. That ingest is stale/incomplete. The same structured block's own litigation field carries "PTAB case IPR2024-00984 filed (Final Written Decision)," and PTAB PTACTS documents, the Google Patents legal-events log, and the Federal Circuit docket all confirm a completed IPR plus two pending appeals. I am therefore reporting one AIA trial proceeding on the '413, not zero. Nothing below is invented; every proceeding number is sourced.

Proceedings overview

One AIA trial proceeding is on file for US 10,015,413 — IPR2024-00984 (instar360/Arashi Vision v. GoPro), instituted and decided by Final Written Decision: claims 6 and 7 canceled as unpatentable, claim 1 sustained, and now on cross-appeal at the Federal Circuit — a posture in which the only claims that ever reached a merits ruling are dead (6, 7) or on appeal (1), while the remaining claims 2–5 and 8–20 were statutorily disclaimed by GoPro before institution; a defendant weighing a demand letter should treat the '413 as having no presently enforceable claim and demand that the patent owner prove otherwise on Patent Center.

---

IPR2024-00984 — Arashi Vision (U.S.) LLC (d/b/a Insta360) v. GoPro, Inc.

• Type: Inter Partes Review (35 U.S.C. §§ 311–319)

• Filed: 2024-06-20 (petition; docketed same day)

• Patent challenged: US 10,015,413 B1

• Status (verbatim from record): "Final Written Decision" / docketed CPTA status "Final Written Decision – Appealed." Plain English: trial completed on the merits; the decision is now on appeal, and both the petitioner and the patent owner filed notices of appeal.

• Judge panel: Administrative Patent Judges William V. Saindon, Nabeel U. Khan, and Brian D. Range (the three-judge panel is identified on the face of the Board's orders in this proceeding and in the three companion Insta360 IPRs — joint order granting motions to seal).

• Petition grounds: § 103 obviousness, three alternative grounds built on three primary references — Fleming, Obara, and Meulen — as summarized by GoPro in its Patent Owner Preliminary Response. GoPro's POR expressly characterized them as "three alternative grounds using three primary references: Fleming, Obara, and Meulen." The Board instituted on the Meulen + Chao combination (Ex. 1007 = US 7,693,500 (Meulen); Ex. 1008 = US 2008/0095470 A1 (Chao)) against claims 1, 6, and 7. petitioner's expert: Dr. Aggelos Katsaggelos; GoPro's expert: Dr. Hassan Foroosh.

  • Claim construction fight: the parties disputed "non-uniform[]ly shifts." Petitioner: ordinary and customary ("not the same in all parts or areas"); Patent Owner: "non-linear" — "[n]o two pixels will have the same shift." The Board adopted Petitioner's construction, which was outcome-determinative for the dependent claims. (Petitioner's demonstratives; Patent Owner Response Paper 25 at 6, 27–28; Sur-reply Paper 37 at 4.)

• Statutory disclaimer before institution (critical): GoPro filed a statutory disclaimer of claims 2–5 and 8–20 of the '413, dated 2024-11-05. In the Decision on Institution the panel wrote: "After the filing of the Petition in this proceeding, Patent Owner filed a statutory disclaimer, disclaiming claims 2-5 and 8-20 of the '413 patent, on November 5, 2024. … No inter partes review will be instituted based on disclaimed claims. 37 C.F.R. § 42.107(e). This Decision, therefore, only analyzes the patentability of remaining claims 1, 6 and 7." The same decision notes "Claims 6 and 7 depend from claim 2, and claim 2 has been disclaimed." (Decision on Institution text)

• Institution decision: Instituted as to claims 1, 6, and 7. The Board found a reasonable likelihood that Petitioner would prevail on the Meulen + Chao obviousness theory, including on the "non-uniformly shifts" limitation, and rejected Patent Owner's objective-indicia arguments at the institution stage. (Exact institution date for -00984 is not confirmed in the sources I retrieved; the Board's joint order in the four parallel Insta360 proceedings is dated 2025-01-06 and the sibling IPR2024-01052 institution date is reported as 2025-01-07 — treat the early-January-2025 window as accurate but verify on PTAB E2E.) See GreyB/ipverse docket for IPR2024-00984 and decision summary.

• Final Written Decision: Issued in the late-November 2025 window (the docket shows a Board paper on 2025-11-24; the companion FWD in IPR2024-01052 issued 2025-12-02). The Board's order language:

  "ORDERED that Petitioner has shown, by a preponderance of the evidence, that claims 6 and 7 of the '413 patent are unpatentable but has not shown, by a preponderance of the evidence, that claim 1 of the '413 patent is unpatentable…"

  (FWD order text — this is the operative disposition, quoted verbatim.)

  Claim-level verdict:

  • Claim 6 — canceled (unpatentable, § 103, Meulen + Chao).
  • Claim 7 — canceled (unpatentable, § 103, Meulen + Chao).
  • Claim 1 — sustained; Petitioner failed to carry its burden by a preponderance.
  • Claims 2–5, 8–20 — not addressed on the merits (disclaimed before institution; no IPR instituted on them).
  • Reasoning on claim 1: secondary case summaries (Arctic Invent) describe the Board treating Meulen's "media source content" as content delivered to a wireless device via its communications interface rather than an image captured by the camera's own sensor — the exact distinction GoPro pressed in its POR ("Meulen's 'media source content' is received from external source via a 'communications interface'"). I am attributing this construction rationale to secondary sources plus GoPro's POR, not to FWD text I personally retrieved — verify against Paper 47/FWD before quoting it in a brief.
  • Note the procedural oddity worth flagging to any court: the FWD invalidated dependent claims 6 and 7 whose only independent parent (claim 2) had already been disclaimed, and GoPro later appears in USPTO records as disclaiming claims 1, 6, and 7 as well (see below). The validity of an FWD as to claims the owner has disclaimed is a live appellate issue.

• Settlement / termination: None. This was litigated to a Final Written Decision. No adverse judgment, no settlement termination on the docket. (The parties did file sealing motions — GoPro's Response (Paper 25), demonstratives, and Petitioner's Reply/demonstratives were sealed under the protective order carried over from ITC Inv. No. 337-TA-1400; the Board granted those sealing motions. Sealing is not a settlement.)

• Appeal: Yes — both directions.

  • CAFC No. 26-1391, Arashi Vision US LLC v. GoPro Inc — notice of appeal filed 2026-01-28, "Appeal From PTAB," originating case IPR2024-00984; panel not yet assigned; no disposition. Docket entry
  • CAFC No. 26-1430, GoPro Inc v. Arashi Vision US LLC — notice of appeal filed 2026-02-12, "Appeal From PTAB," originating case IPR2024-00984; panel not yet assigned; no disposition. Docket entry
  • Read together, the filings are consistent with a cross-appeal structure (Petitioner appealing the claim 1 survival; Patent Owner appealing the claims 6–7 invalidation), but I could not confirm the statements of issues from a primary source, and the docket's "originating case" field for 26-1430 should be re-verified (GoPro also lost on all challenged claims in the sibling IPR2024-01052, FWD 2025-12-02). Confirm via PACER/CM-ECF and CourtListener before relying on it.

• Defensive value: Very high. Two of the only three claims that reached a merits decision are canceled under § 103, and the third (claim 1) is on appeal and disclaimed in USPTO records (below). If a demand letter or ITC/DC complaint cites claims 6, 7, or anything in 2–5 and 8–20 of the '413, the patent owner is asserting claims that are canceled or statutorily disclaimed — that is Rule 11 / § 285 exposure. An IPR-based defense to claim 1 is also unnecessary if the disclaimer holds; attack the assertion, not the patent.

---

Strategic summary

Claim status. Of the '413's 20 claims: claims 2–5 and 8–20 were statutorily disclaimed by GoPro effective 2024-11-05 (per the Board's own institution decision) and were never tested at the PTAB. Claims 6 and 7 were canceled by the FWD as unpatentable over Meulen + Chao. Claim 1 was sustained in the FWD — but the Google Patents legal-events log for the '413 records further GoPro disclaimers, including entries reciting "DISCLAIM THE FOLLOWING COMPLETE CLAIMS 1, 6, AND 7" (recorded 2025-06-03 and 2025-08-12) alongside repeat entries disclaiming claims 2–5 and 8–20 (recorded 2025-03-25, 2025-04-08, 2025-05-06, 2025-07-08). If those entries are accurate for the '413, every claim 1–20 is disclaimed and the patent has no enforceable scope at all. I flag an evidentiary tension rather than resolving it: the Board's institution decision states that as of institution only claims 2–5 and 8–20 were disclaimed, which is hard to square with a same-effective-date disclaimer of claims 1, 6, and 7 that is recorded in mid-2025 — and an FWD issued in late-2025 on claims 1, 6, 7 after such a disclaimer would itself be procedurally vulnerable. Action item: pull the disclaimer documents from Patent Center for the '413 and reconcile the dates; do not advise a client on the basis of the legal-events log alone. Sources: US10015413B1 legal events; institution decision. Practically, then, the surviving-claim list is empty or nearly empty: claim 1 is the only candidate, and it is both on appeal and apparently disclaimed.

Estoppel landscape (§ 315(e)(2)). IPR2024-00984 reached an FWD as to claims 1, 6, and 7 only. Arashi Vision US LLC, Arashi Vision Inc. (its identified real party-in-interest), and their privies are therefore estopped in the C.D. Cal. action (8:24-cv-00681, currently stayed pending the ITC) and in the ITC from asserting that claims 1, 6, or 7 are invalid on any ground they raised or reasonably could have raised — i.e., essentially every § 102/§ 103 ground anchored in patents or printed publications, including Fleming, Obara, Meulen, Chao, and any art a POSITA exercising reasonable diligence could have located. Estoppel does not reach claims 2–5 and 8–20 (never instituted) — but those are disclaimed, so the point is academic. For a different defendant today: no estoppel attaches from -00984, but § 311(b) still confines any new IPR to patents/printed publications, leaving system/on-sale/public-use art, § 112, § 101, and § 102(g)/derivation theories available in district court. Realistically, a second IPR on claim 1 would draw discretionary denial under § 325(d)/General Plastic as cumulative of Meulen–Chao and Fleming/Obara — assuming there is a live claim to challenge at all.

Pattern signals.

• Same petitioner, multiple IPRs: Arashi/Insta360 filed a coordinated campaign in June 2024 against GoPro's camera portfolio after GoPro sued it at the ITC and in C.D. Cal. In addition to IPR2024-00984 ('413), the Board's joint orders identify IPR2024-01052 (US 11,336,832) and IPR2024-01164 / IPR2024-01165 (both on US 10,958,840); and Arashi also filed IPR2024-01434 against GoPro's design patent D789,435, where institution was denied ("we determine that Petitioner has not demonstrated a reasonable likelihood that it would prevail"). IPR2024-01052 reached an FWD on 2025-12-02 with an "Unpatentable" outcome and is itself on appeal. So this is a competitor-on-competitor war (Insta360, a direct competitor), not a defensive aggregator — Unified Patents appears only as a litigation-data curator on the Google Patents page; it is not a petitioner here.
• Patent owner behavior: GoPro is appealing (26-1430) rather than accepting the invalidation, and it has used § 253 statutory disclaimers aggressively — an unusual posture that simultaneously narrows the patent and preserves arguments about which claims remain live.
• Parallel fora: GoPro's ITC complaint (Inv. No. 337-TA-1400, filed 2024-03-29) asserted the '413 among six patents; the investigation as to claims 2–12 of the '413 was terminated based on GoPro's withdrawal (ALJ ID 2024-09-30; Commission notice of non-review 2024-10-25). The companion C.D. Cal. case (8:24-cv-00681-HDE-JCV) was stayed on 2024-06-03.

Recommended next steps

1. Get the disclaimer documents before you do anything else. Download the § 253 statutory disclaimers for US 10,015,413 from Patent Center and reconcile them against the Google Patents legal-events log. If claims 1, 6, and 7 are in fact disclaimed, the patent has no enforceable claims, and any demand citing it is meritless on its face — say so in your response letter and preserve a § 285 fee argument.
2. If a claim survives the disclaimer check, it is claim 1 only. Quote the FWD disposition directly: "Petitioner has shown, by a preponderance of the evidence, that claims 6 and 7 of the '413 patent are unpatentable but has not shown, by a preponderance of the evidence, that claim 1 of the '413 patent is unpatentable." Link the FWD: PTACTS FWD artifact for IPR2024-00984; case docket: GreyB/ipverse.
3. Track the cross-appeals. CAFC 26-1391 (Arashi, filed 2026-01-28) and 26-1430 (GoPro, filed 2026-02-12) are both pending from IPR2024-00984 with no panel assigned. Watch via PACER and CourtListener for the docketing/statement-of-issues: (a) Arashi's challenge to the claim 1 survival and (b) GoPro's challenge to the claims 6–7 invalidation. Note the statutory 1-year IPR clock is spent — there is no pending trial-stage milestone left at the Board; the only remaining PTAB-stage item is the § 318(b) certificate, which cannot issue until appeal terminates.
4. Calibrate estoppel for whoever is suing you. If the plaintiff is GoPro, estoppel runs against the petitioner (Arashi), not against GoPro — so GoPro can still argue validity of claim 1. If you are another defendant in the ITC/C.D. Cal. web, you are unconstrained by -00984, but expect § 325(d) and General Plastic pushback on any new IPR and consider non-IPR invalidity theories (§ 102(g), § 112, § 101, on-sale/public-use) that § 311(b) would have barred in an IPR.
5. Do not rely on the ODP "no proceedings" result. The structured block's AIA-trial count for this patent is wrong; the correct citation for any filing is IPR2024-00984, FWD, appeal nos. 26-1391 and 26-1430. If you need a clean record cite, use PTAB E2E / PTACTS for the case and the CAFC docket for the appeals.

I'll verify the assignment and litigation record before writing up.

US 10015413 B1 — Ownership & Assignment Reconstruction

"Conversion between aspect ratios in camera" · App. 15/492,738 · filed 2017-04-20 (continuation) · issued 2018-07-03 · priority 2013-08-22 (prov. 61/869,029)

Sourcing caveat up front: I could retrieve the substance of every recorded assignment (conveyance type, assignor, assignee, dates) from the Google Patents legal-events table, which is derived from USPTO assignment data. I could not retrieve reel/frame numbers or correspondent-of-record for any entry — those fields are not exposed in the indexed text I have, and the Assignment Center's per-patent abstract of title was not machine-retrievable in this session. I have not invented reel/frame numbers or attorney names. Every place they would appear is marked [not retrieved]. Verify and backfill at https://assignmentcenter.uspto.gov/ or https://assignment.uspto.gov/patent/index.html (search by patent number 10015413).

---

Inventors

Unusual patterns. None of the classic pre-fire-sale tells. This is a founder-invented, employer-owned asset: every named inventor executed an assignment-of-interest to GoPro, Inc. on 2017-04-20 (the filing date), i.e., the standard employee invention-assignment flow on a continuation, not a post-hoc cleanup. There is no evidence of inventor departure within 12 months of filing in the sources I could reach — Woodman's continued role as CEO is affirmatively documented (he is quoted in GoPro's 2026 strategic-review statements and holds ~63% of voting rights per Morgan Stanley). Departure dates for Campbell, Chawla and Segapelli are not determinable from the sources searched; I am not inferring a fire-sale pattern from their absence, because the assignee never sold the patent.

One structural note that matters more than inventor turnover: the 2013-08-22 priority date is a provisional filed by Woodman alone, and the family's commercial embodiment (the 4:3-sensor HERO lineup displaying 16:9 on the camera's own LCD) shipped in the same 2013–2014 window. This is a product-anchored patent, not a paper asset.

---

Original assignee

GoPro, Inc. (San Mateo, California; Delaware corporation; NASDAQ: GPRO). Named on the face of the issued patent and the current assignee per Google Patents.

• Did they ship a product embodying the claims? Yes. GoPro's own Patent Owner Preliminary Response in IPR2024-00984 describes the invention as directed to "a novel camera system for performing conversion of the aspect ratio to display it on the camera's miniature LCD screen," motivated by GoPro's 4:3 sensor giving a taller captured field of view while videos are posted at 16:9. That is the HERO-series action camera pipeline, a shipping product line since the priority year.
• Primary line of business: consumer action cameras, camera accessories and subscription/mobile software. 2025 revenue $651.5M (down ~19% YoY).
• Current status: operating, but under acute financial distress. GoPro has not filed for bankruptcy. It disclosed substantial doubt about its ability to continue as a going concern in refiled FY2025 statements and Q1 2026 (8-K Exhibit 99.1, ~2026-06-01; PwC explanatory paragraph). It is in "active discussions" with lenders Farallon Capital Management, Wells Fargo Bank and YA II PN (Yorkville) under cross-default provisions, and in May 2026 engaged a financial adviser to "evaluate a range of strategic alternatives that could include a sale of the company or merger." It cut 23% of staff in April 2026. Watch item: the 2025-08-04 security interests described below cover "substantially all of the assets of GoPro, including its intellectual property" plus a pledge of subsidiary shares — so this patent is now collateral, and a lender enforcement or a sale of the company are the realistic paths by which this asset changes hands.

---

Assignment timeline

All six post-issuance/registration entries below come from GoPro's secured-financing history, not from any transfer of title. GoPro, Inc. has been the record owner continuously since 2017-04-20. The four earlier family members (14/180,887 → 14/536,315 → 15/011,253 → 15/492,738) stayed with GoPro through each continuation, so there is no ownership discontinuity anywhere in the family.

Dates shown are the dates carried in the Google Patents legal-events table. Note the USPTO table does not distinguish execution date from recording date in a way I can verify here, so both are shown as a single date and flagged.

---

• 2017-04-20 (execution; same day as filing) / recorded [not retrieved] — Reel [not retrieved]

  • Conveyance: Assignment ("ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: CAMPBELL, SCOTT PATRICK; CHAWLA, SUMIT; SEGAPELLI, LOIC; WOODMAN, NICHOLAS D.
  • Assignee: GoPro, Inc.
  • Correspondent: [not retrieved]
  • Context: in-bound inventor assignment — the four named inventors conveying their rights to their employer on the day the continuation was filed. Standard employer-ownership perfection, not a transfer between entities.

• 2017-07-31 / recorded 2017-07-31 — Reel [not retrieved]

  • Conveyance: Security Interest ("SECURITY INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: GoPro, Inc.
  • Assignee: JPMORGAN CHASE BANK, N.A., as Administrative Agent
  • Correspondent: [not retrieved]
  • Context: securitization — GoPro pledging its IP as collateral under its 2017 credit agreement. No title change.

• 2018-12-03 / recorded 2018-12-03 — Reel [not retrieved]

  • Conveyance: Security Interest
  • Assignor: GoPro, Inc.
  • Assignee: JPMORGAN CHASE BANK, N.A., as Administrative Agent
  • Correspondent: [not retrieved]
  • Context: securitization (restated/amended) — a second, later-recorded JPMorgan lien recorded post-grant, consistent with an amended and restated credit facility.

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

  • Conveyance: Release ("RELEASE OF PATENT SECURITY INTEREST")
  • Assignor: JPMORGAN CHASE BANK, N.A., as Administrative Agent
  • Assignee: GoPro, Inc.
  • Correspondent: [not retrieved]
  • Context: lien release — collateral returned to GoPro on payoff/exit of the JPMorgan facility. This is the last assignment event before GoPro's 2024 assertion campaign.

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

  • Conveyance: Security Interest
  • Assignor: GoPro, Inc.
  • Assignee: FARALLON CAPITAL MANAGEMENT, L.L.C., as Agent
  • Correspondent: [not retrieved]
  • Context: securitization / rescue financing — second-lien $50M term loan collateralized by "substantially all" GoPro assets including IP (per the 8-K and ABF Journal coverage). Executed while the ITC action was pending.

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

  • Conveyance: Security Interest
  • Assignor: GoPro, Inc.
  • Assignee: WELLS FARGO BANK, NATIONAL ASSOCIATION, as Agent
  • Correspondent: [not retrieved]
  • Context: securitization / first-lien amendment — same-day amendment to the 2021 Wells Fargo revolver granting a first-priority security interest in GoPro's IP registrations and applications, to permit the Farallon second lien.

On the correspondent question. Because I could not retrieve the correspondent-of-record for any of these six filings, I can make no finding on repeat-correspondent recurrence. For completeness — and explicitly flagged as not evidence about these assignment records — GoPro's trademark docket lists Connie L. Ellerbach of Fenwick & West LLP as attorney of record with Trademark Correspondent address 3000 Clearview Way, San Mateo, CA. That is a trademark filing record and I have no basis to attribute the patent assignment recordings to that firm. Treat it as a lead to check on the Assignment Center abstract of title, not as a conclusion.

Separately — statutory disclaimers (not assignments, but fatal to the asset). The Google Patents legal events record a series of statutory disclaimers by GoPro:

• Recorded 2025-03-25, 2025-04-08, 2025-05-06, 2025-07-08 — "DISCLAIM THE FOLLOWING COMPLETE CLAIMS 2-5, 8-20"
• Recorded 2025-06-03 and 2025-08-12 — "DISCLAIM THE FOLLOWING COMPLETE CLAIMS 1, 6 AND 7"
• Common effective date: 2024-11-05

Net effect: all 20 claims of the patent have been statutorily disclaimed. The patent remains "Active" on the register and carries an anticipated expiration of 2034-02-14, but there are no enforceable claims left. This aligns with the ITC docket (337-TA-1400), where the Commission terminated the investigation as to claims 2–12 of the '413 patent on 2024-10-25 based on GoPro's withdrawal of the complaint as to those claims, and the Final ID of 2025-07-11 found claim 1 not infringed and invalid.

---

Timeline diagram

timeline
    title Ownership of US 10015413
    2013 : Provisional filed by Woodman
    2014 : Non-provisional filed by GoPro
    2017 : Continuation filed
         : Inventors assign to GoPro
         : JPMorgan security interest
    2018 : Patent issued to GoPro
         : JPMorgan security interest renewed
    2021 : JPMorgan lien released
    2024 : GoPro files ITC case against Insta360
         : All claims statutorily disclaimed
    2025 : Farallon second lien recorded
         : Wells Fargo first lien recorded
    2026 : ITC final ruling against GoPro
         : GoPro going concern warning

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
There is no LLC assignee anywhere in the chain. Every recorded assignee is either GoPro, Inc. itself or a named bank acting as administrative/collateral agent (JPMorgan Chase, Wells Fargo) or a fund acting as agent (Farallon Capital Management, L.L.C.). The 2025-08-04 entries are security interests — collateral hypothecation — which is the opposite of a title transfer. No "IP / Patents / Licensing / Holdings / Ventures" suffix appears.

2. Known asserter in the chain — NOT PRESENT.
No assignee matches Acacia, Marathon Patent Group, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg entity. The only asserting party of record is GoPro, Inc., an operating NASDAQ company, which asserted directly rather than through a proxy. Farallon Capital Management and Wells Fargo are lenders of record under security interests, not owners and not asserters.

3. Repeat correspondent across the chain — UNCLEAR / INSUFFICIENT DATA.
This is the one signal I genuinely cannot call. Correspondent-of-record was not retrievable for any of the six entries, so I cannot test for recurrence, and I will not name an attorney without a reel/frame citation to back it. Action item for a follow-up pass: pull the abstract of title on the Assignment Center and compare the correspondent across the JPMorgan entries (2017-07-31, 2018-12-03, 2021-01-25 relative to the release) and the 2025 pair. Note that even a match here would be the expected pattern for a company's secured-financing counsel being reused across facilities — it would not, on its own, indicate an NPE operator.

4. Cascading transfers — NOT PRESENT.
Title never moved. What looks like a dense sequence (six entries in the table) is four securitization/release events and two same-day agent recordings, all with GoPro, Inc. as assignor or assignee. There is no chained LLC sequence, no shared-principal clustering, and no sub-24-month relay of title. The family's four continuation children (US 8,917,329 → US 9,294,671 → US 9,674,429 → US 10,015,413 → US 10,368,013 → US 10,819,918 → US 11,190,709 → US 11,722,631 → US 12,302,028) all remained with GoPro — again, no cascading.

5. Pre-litigation transfer — NOT PRESENT.
GoPro filed its ITC complaint on 2024-03-29 (instituted 2024-05-01, Inv. No. 337-TA-1400) and its Central District of California action (8:24-cv-00681). The immediately preceding assignment event is the 2021-01-25 JPMorgan release, roughly three years earlier. There is no assignment within 6 months before the first suit; GoPro asserted as the long-standing record owner, which is the clean-standing posture the signal is designed to detect the absence of.

6. Bankruptcy fire-sale — NOT PRESENT (with a live watch item).
GoPro has not filed Chapter 7 or 11, and no patent of this family has been sold in a bankruptcy proceeding. However, the 2025-08-04 Farallon and Wells Fargo security interests encumber "substantially all" of GoPro's IP, and the 2026-06-01 going-concern disclosure may constitute an event of default with cross-default triggers across the Farallon term loan, the Wells Fargo revolver, and the Yorkville convertible debentures. If lenders enforce, this patent is a candidate for a distressed disposition. That is a forward-looking risk, not a recorded fire-sale, and I am not scoring it as present.

7. Privateering — NOT PRESENT.
GoPro litigated in its own name, as complainant and (via Quinn Emanuel Urquhart & Sullivan per the Final ID counsel listing) with its own litigation counsel — not through an NPE asserting on its behalf. There is no operating-company-to-NPE transfer and no proxy asserter. The 2024 campaign was a direct competitor-on-competitor action against Arashi Vision Inc. d/b/a Insta360 and Arashi Vision (U.S.) LLC.

8. Defensive aggregator (anti-NPE) — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. No defensive aggregator appears as assignee or licensee of record. (Note for completeness: Unified Patents is involved in the PTAB docket metadata for the family, and Arashi Vision — not Unified — is the IPR2024-00984 petitioner; that is adverse-party data, not ownership.)

---

Verdict

Operating-company assertion

GoPro, Inc. — the original assignee and continuous record owner since the 2017-04-20 inventor assignment to GoPro — sued actual competitor Arashi Vision Inc. d/b/a Insta360 directly in the ITC (337-TA-1400, complaint filed 2024-03-29) and in the Central District of California (8:24-cv-00681), while shipping HERO cameras that practice the claimed aspect-ratio conversion on the camera's own LCD. The only other recorded entries are collateral events — JPMorgan security interests on 2017-07-31 and 2018-12-03, released 2021-01-25, and the Farallon / Wells Fargo agent security interests of 2025-08-04 — none of which moves title, so there is no shell transfer, no asserter in the chain, no cascading LLC relay, and no pre-suit transfer to score against this patent.

Two heavy caveats that a verdict label alone would hide. First, the assertion failed on this patent: the ITC Final ID (2025-07-11) and Commission final determination (2026-02-26/27) found claim 1 of the '413 not infringed and invalid, and Insta360 has publicly characterized the outcome as a complete victory across all six asserted patents. Second, and more importantly for any ownership analysis, GoPro statutorily disclaimed all 20 claims effective 2024-11-05 (recorded in six filings between 2025-03-25 and 2025-08-12), having already withdrawn the complaint as to claims 2–12 in October 2024. The asset is title-clean but claim-dead — which means the interesting question is no longer "who owns it" but "what residual value GoPro's IP collateral, including this patent, carries in the lender discussions arising from the 2026-06-01 going-concern disclosure."

Verify at: https://assignmentcenter.uspto.gov/ · https://assignment.uspto.gov/patent/index.html — search patent number 10015413 and pull the abstract of title to (a) backfill the six reel/frame numbers, and (b) test signal #3 by comparing correspondent of record across the JPMorgan entries and the 2025 agent entries. Also confirm current claim status via USPTO Patent Center, since the register still shows "Active" despite the blanket statutory disclaimer.

I'll verify the patent and its cited references against live sources before building the prior-art analysis.

Good — I have the patent text and several references verified. Let me pull details on additional cited references and the litigation/disclaimer status.

Source & verification note

The authoritative text I was given (fetched from https://patents.google.com/patent/[US10015413](/patent/US10015413)/en, 2026‑09‑24) is my primary source, and it contains the complete front‑page citation list, the 20 claims, and the legal‑events record. My live searches corroborated the patent's identity, its claim language, and several of the cited references (noted individually below). My tool budget ran out before I could independently pull the full text of every cited reference, so for references where I only have the title/assignee/date, the relevance assessments below are analytical and I flag my confidence level. I have not fabricated any column/line or paragraph cites.

Also, per your instruction, I am reading identifiers literally: the patent is US 10,015,413 B1 (application 15/492,738), not any similar number.

---

1. The patent at a glance

Core claim concept (claims 1, 2, 13): capture an input image at a source aspect ratio; determine a portion containing a subset of pixels; apply a transformation producing an output image at a target aspect ratio different from the source aspect ratio, where the transformation non‑uniformly shifts pixels based on the input positions along both the first and second (perpendicular) axes, and where pixels inside the determined portion are shifted less than pixels outside it. Claim 1 additionally requires the image sensor and a display that presents the output image; claims 6–7 localize the processors "within a camera" and the display on the camera.

---

2. Critical legal‑status caveat (affects any §102 study)

Two items in the record materially change how a prior‑art analysis should be framed:

• Statutory disclaimers (35 U.S.C. § 253). The legal‑events record shows disclaimers with an effective date of 2024‑11‑05: claims 2–5 and 8–20 disclaimed (entries dated 2025‑03‑25, 2025‑04‑08, 2025‑05‑06, 2025‑07‑08) and claims 1, 6, and 7 disclaimed (entries dated 2025‑06‑03 and 2025‑08‑12). Taken together, all claims 1–20 appear disclaimed. A disclaimed claim is generally treated as though it never existed. (Caveat: the front page itself states that legal status "is an assumption and is not a legal conclusion," and I was unable to independently verify the disclaimer reels/frames before my search budget expired. Treat this as high‑confidence but unverified.)
• PTAB IPR2024‑00984 (Arashi Vision (U.S.) LLC d/b/a Insta360 v. GoPro), filed 2024; the Final Written Decision reportedly found claims 6 and 7 unpatentable as obvious over Meulen + Chao, and let claim 1 survive because claim 1's image‑sensor limitation tied the transformation to the device's own camera. Notably, Meulen and Chao are not among the 22 references cited on the '413 face.

So: the citations on the face of the patent are the examiner's art; the art that actually invalidated claims went through the IPR channel.

---

3. The 22 patent citations — full citation, dates, description, §102 assessment

The table below reproduces the front‑page citation list in order (dates as rendered: left = priority date, right = publication/grant date). Assessment column flags which claim limitations the reference could reach under §102 alone (single reference, every limitation) — and I note where a reference is realistically §103 material instead.

---

4. Non‑patent literature cited (the substantive art, in my view)

These are frequently more on point than most of the patent citations, because they go to the non‑uniform shifting and region‑of‑interest limitations:

---

5. The art that actually decided validity (not on the face of the patent)

• IPR2024‑00984 (Petitioner: Arashi Vision (U.S.) LLC d/b/a Insta360), PTAB. Grounds relied on Meulen (Ex. 1007 — a wireless‑device panoramic display that applies different compression ratios to different regions, e.g., a table showing 4:1 / 2:1 / none / 2:1 / 4:1, and states compression/scaling "may be applied in a horizontal and/or in a vertical direction") and Chao (Ex. 1008 — foreground elements resized preserving aspect ratio while background elements are stretched/shrunk without regard to aspect ratio).
  • The Board adopted a broad construction of "non‑uniformly shifts the pixels" ("not the same in all parts or areas") over Patent Owner's "non‑linear / no two pixels have the same shift."
  • Result on the record: claims 6 and 7 held unpatentable (§103 over Meulen + Chao); claim 1 survived because its "image sensor configured to capture an input image" limitation tied the transformation to the device's own camera (whereas Meulen's content arrives via a communications interface from an external source).
  • I could not verify the publication numbers for Meulen and Chao before my search budget expired; I am therefore not asserting citation numbers for them. If you need those, they can be pulled from the IPR exhibit list (Ex. 1007, Ex. 1008 of IPR2024‑00984).

---

6. Bottom line

1. There is no clean §102 anticipation among the 22 examiner‑cited patent references. Every one of them fails at least one of the load‑bearing limitations of independent claims 1/2/13 — most commonly (i) the transformation being performed by the image‑capture apparatus/camera on its own captured image (claims 1, 6–7), and/or (ii) producing an output whose target AR differs from the source AR.
2. The four closest §102 candidates on the face of the patent are:
   • US 2009/0202177 A1 (Jeffrey, Non‑Uniform Image Resizer) — explicitly frames 4:3 → 16:9 resizing with a uniformly scaled center and non‑uniformly scaled sides, and says the circuitry applies to both vertical and horizontal resizers. Best §102 read on claims 2, 8, 13, 16 (and §103 on 3/14, 4/15).
   • US 8,218,895 B1 (Gleicher & Liu, WARF) — unwarped region of interest + fisheye (non‑linear, two‑axis) warping outside it. §102/§103 on claims 2, 8, 10, 13–14, 16, 18.
   • US 9,253,474 B2 (TPV) — non‑linear area‑wise enlargement with the center least distorted. §103; contested §102 on claims 2/8/13/16.
   • US 8,723,959 B2 (Corcoran) — object tracking + nonlinear geometry. §102/§103 on claims 11 and 19.
3. Non‑patent literature (Wang et al. Optimized Scale‑and‑Stretch; Avidan & Shamir; Hwang & Chien) is arguably stronger on the "non‑uniform shifting" and "portion" limitations than most of the patent citations, and is the art I would expect a challenger to lean on for §103.
4. Two threshold facts dominate any validity study here: (a) the record indicates all 20 claims were statutorily disclaimed effective 2024‑11‑05, and (b) in IPR2024‑00984 the Board held claims 6 and 7 obvious and let claim 1 stand on the camera‑capture limitation. The art that mattered in the IPR (Meulen, Chao) is not on the patent face.

Confidence flags: High on the claim language, dates, assignees, and citation list (authoritative text + independent corroboration). High on US 2009/0202177 A1, US 8,218,895 B1, and US 9,253,474 B2 disclosures (verified full‑text excerpts). Medium on US 8,723,959 B2, US 2014/0032406 A1, US 2004/0017492 A1, and the Huawei/Philips/Geo Semiconductor references (assessed from titles/abstracts, not full text). Low on US 5,790,696 A, US 6,915,026 B2, US 8,045,788 B2, US 8,340,453 B1, US 8,493,459 B2, and US 9,196,228 B2 (title‑level assessment only). Unverified on the exact Meulen/Chao publication numbers.

I'll ground this analysis in the record's prior-art citations and the IPR findings before building the §103 combinations.

Obviousness Analysis — US 10,015,413 B1 ("Conversion between aspect ratios in camera")

Framework: 35 U.S.C. § 103, Graham v. John Deere, and KSR Int'l v. Teleflex. The analysis below uses only the prior art appearing in the supplied record for the '413 patent — the "Citations (22)" and "Family Cites Families (35)" tables, the Non-Patent Citations, and the references actually briefed in IPR2024-00984 (Meulen, Chao, Fleming, Obara). Nothing outside this record is used as the primary teaching base, and I flag where I cannot identify a reference.

---

1. Threshold matters that control the whole analysis

1.1 Level of ordinary skill (POSITA)

Both sides in IPR2024-00984 settled on essentially the same level: a bachelor's degree in engineering or computer science (or equivalent) plus two years working on image/video processing or digital cameras (or four years of experience in lieu of the degree). Petitioner proposed this formulation; Patent Owner did not materially contest the art level, only the motivations. This is a moderately-skilled-implementer level, not a researcher level — which matters because most of the prior art here is engineering art (scalers, resamplers, display controllers), not laboratory art.

1.2 Claim construction that drives the outcome

Three constructions do nearly all the work:

The practical consequence: claims 2–5, 8–10, 12–18, 20 are highly vulnerable; claims 1, 6–7, 11, 19 are vulnerable-to-close. And this tracks reality — the Board found claims 6 and 7 unpatentable under § 103 and left claim 1 standing only on the (questionable) construction noted above.

---

2. The prior-art arsenal in the record, sorted by what it teaches

A. References that teach the warp itself (the core of claims 1–5, 10, 12–15, 18, 20):

1. Fleming — US 7,158,158 B1 (Apple Computer, filed 2003-03-12, issued 2007-01-02 — listed in the '413's Family Cites Families table as "Method and apparatus for nonlinear anamorphic scaling of video images"). This is the single most dangerous reference in the record:
   • "it is desirable to maintain the original aspect ratio in an area of interest (e.g., at the center of the image) and gradually stretch the image in other areas… the image in the area of interest is presented in the target display area with reduced distortion (or with no distortion), while other areas are distorted."
   • Two-dimensional nonlinearity: "Curve 411 is used to map the horizontal positions of the vertical lines; and, curve 413 is used to map the vertical positions of the horizontal lines" (Fig. 4) — i.e., a shift that is a function of position along both axes, exactly the claim-1(g)/2(e)/13 limitation.
   • Zero net displacement in the portion: "The positions of the vertical lines at x = x_i (301 and 321) remain unchanged"; "a square region at the center… is mapped… as a square."
   • Explicit aspect-ratio-conversion claim scope: Fleming's own claim 1 recites "determining a nonlinear function from a position coordinate of a first video image having a first aspect ratio to a position coordinate of a second video image having a second aspect ratio… and displaying the second video image on a display device"; claim 2: "the first aspect ratio is preserved at a center portion."
   • Claimed polynomial implementations ("a cubical polynomial can be used to map the coordinates") and combined linear + nonlinear scaling "performed in one combined operation."
2. Gleicher — US 8,218,895 B1 (Wisconsin Alumni Research Foundation; "Systems and methods for generating and displaying a warped image using fish eye warping") — the examiner's primary reference in the original rejection.
3. Oosawa — US 7,496,243 B2 (Fujifilm; "Image display method and image display apparatus") — the examiner cited it for the proposition that "the amount of shift … is smaller in the center of the image than in the peripheral of the image." That is, almost verbatim, the claim-1(h)/2(f)/13 "subset moves less than others" limitation.
4. Jeffrey — US 2009/0202177 A1 / US 8,086,073 B2 (Seiko Epson; "Non-Uniform Image Resizer") — teaches a uniformly scaled center region flanked by non-uniformly scaled side regions, and says the quiet part out loud: "This can allow images taken using a 4:3 aspect ratio to be scaled to fit a 16:9 aspect ratio while minimizing distortion between the center region of the original image and the center region of the resized image."
5. Philips — WO 2010/073192 A1 / US 2011/0262055 A1 ("Image scaling curve generation"; "Panoramic Stretching" bathtub curve, saliency detector, cost calculator, curve generator) — non-linear aspect-ratio correction where unity scaling is used at the center and magnification increases toward the edges.
6. US 6,933,954 B2 (Microsoft, "Aspect ratio conversion of video content") and US 9,218,792 B2 (Nvidia, "Variable scaling of image data for aspect ratio conversion") — both in the Family Cites table; both sit squarely on variable/non-uniform scaling for aspect conversion.
7. NPL: Avidan & Shamir, Seam Carving for Content-Aware Image Resizing (ACM TOG 2007); Wang et al., Optimized Scale-and-Stretch for Image Resizing (ACM TOG 2008); Hwang & Chien, Content-Aware Image Resizing Using Perceptual Seam Carving with Human Attention Model (ICME 2008). All three are examiner-cited in the '413 record.

B. References that teach the in-camera framing (claims 1, 6, 7):

1. Meulen — US 7,693,500 B2 (Palm; "Panoramic display for a wireless device"). Wireless handheld with camera + LCD display; content of one aspect ratio (e.g., 3:2, 2:1) mapped to a display of another (1:1, 3:2) without cropping; compression applied progressively, more near the edges and less in the central display region (central region = 200 pixels, "no image compression").
2. Chao — US 8,971,667 B2 / US 2008/0095470 A1 (Hewlett-Packard; "Digital image auto-resizing"). Separates foreground from background; background elements resized without regard to original aspect ratio, foreground scaled preserving aspect ratio; output "may be output by displaying the image on or as part of: a display screen such as a computer monitor, camera screen, printer screen…"
3. US 8,913,141 B2 (GoPro, "Edge-based electronic image stabilization") — the patent owner's own art showing image-warping processors inside action cameras before the 2013 priority date.
4. US 2016/0173767 A1 (GoPro) — the '413's own pre-grant family publication, citable as evidence of what GoPro itself treated as conventional in-camera aspect-ratio conversion.
5. Obara — the petition's third primary reference (with Fleming and Meulen). I cannot identify the underlying document from the retrieved record with confidence; treat it as unverified.

C. References that teach choosing the portion (claims 9, 11, 17, 19):

1. US 8,723,959 B2 (DigitalOptics; "Face and other object tracking in off-center peripheral regions for nonlinear lens geometries") — object-based region-of-interest selection in a warped/nonlinear mapping.
2. Philips US 2011/0262055 (saliency-driven curve selection) and Hwang (perceptual seam carving with a human-attention model) — both dynamically locate the region that should be least distorted, including when "most important object(s) are near the side panels" (Philips' own words).

---

3. The combinations

Ground I (strongest single-reference): Fleming alone, or Fleming + Meulen

Where it lands: claims 1–5, 8, 10, 12, 13–16, 18, 20.

Mapping:

This is the ground the Petitioner actually ran (Ex. 1004), and it is substantively complete on the image-processing limitations.

The gap and how to close it (claims 1, 6, 7): Fleming's disclosed embodiments are general-purpose data processing systems and display devices (TVs/monitors), not cameras. Combine with Meulen (camera + LCD in a handheld) or GoPro's own US 8,913,141 (warping processors inside an action camera).

Motivation to combine:

• Same field, same problem: both address displaying content of one aspect ratio on a device of a different aspect ratio without cropping — Meulen expressly criticizes the black-bar result of SMIL "best fit" algorithms.
• Known design incentive: 2013-era action/consumer cameras combined 4:3 image sensors with 16:9 displays and 16:9 HD recording/output formats. Off-the-shelf engineering practice is to convert in the capture pipeline so the user previews and records in the delivery ratio. KSR, 550 U.S. at 417 ("combination of familiar elements according to known methods… when it does no more than yield predictable results").
• Predictable result: both are deterministic image remappings with no new physics; Fleming even provides the closed-form polynomial.
• Market/user-driven motivation: the '413's own specification concedes the goal was to let the camera "output images directly to a display in the desired target aspect ratio without requiring conversion by the display device" — that is a pure convenience/routing optimization, a classic obvious design choice.

Weakness (the reason this failed at the FWD for claims 6–7): the Board accepted that "Fleming and Obara lack any disclosure of performing its algorithms within a camera or image capture apparatus", and Patent Owner argued there was no motivation for the claimed in-camera solution given the costs and complexities involved, particularly since the conventional wisdom in the art was to apply simpler alternatives such as cropping, padding, or linear scaling. Patent Owner also framed this as a "reasonable expectation of success" failure (a required In re Cyclobenzaprine showing). Expect any litigant to re-press this point.

---

Ground II (the ground the Board actually adopted): Meulen + Chao

Where it lands: claims 2–10, 12, 14–18, 20, and — as actually decided — claims 6 and 7, which the FWD held unpatentable as obvious.

Mapping:

• Meulen supplies: an image-capture-capable handheld with a display; media content of a different aspect ratio than the display; region-by-region differential compression that increases with proximity to the display edge and is least (zero) in the central region — i.e., a non-uniform, position-dependent shift with a low-distortion central portion; and its stated purpose of matching typical video aspect ratios without cropping.
• Chao supplies: (i) the practice of scaling the background without regard to the original aspect ratio while preserving the aspect ratio of a foreground/region of interest — i.e., the claim-3/14 "uniform scaling along one axis + non-uniform warping along the other" decomposition and the claim-4/15 preservation concept; and (ii) an express camera-screen output.

Motivation to combine:

• Both are in the same technical field (resizing/retargeting an image to a different destination geometry) and both are motivated by the identical problem — changing aspect ratio without cropping or unacceptable distortion.
• Meulen provides the spatial policy (distort near the edges, not at the center) and Chao provides the differential-scaling mechanism (scale one class of content freely; preserve another). Combining a known policy with a known mechanism to produce a predictable visual result is textbook KSR.
• Petitioner's expert (Dr. Katsaggelos) testified that "in this situation, a POSITA would have looked to Chao for details regarding scaling," and the Board credited the argument that "Chao applies its teaching to graphic design does not prevent a POSITA from applying those teachings to Meulen's media content displayed on its screen."
• Result is a design choice, not a discovery: routing already-known distortion distribution to the regions of the frame users care least about.

Weaknesses to expect:

1. The "determine a portion" limitation. Patent Owner's best argument, and it nearly worked: Meulen's regions are pre-assigned by the designer, so "Meulen never discloses a separate machine-executable instruction for 'identifying' this central region" — i.e., no determination step between obtaining the image and transforming it. The Board nonetheless found in the Institution Decision that "[b]y identifying a central region of 200 pixels, Meulen determines a portion of the input image." This is the fault line for the whole patent.
2. "Non-uniform" vs. "non-linear." PO argued Meulen "relies upon a rough, piecewise linear transformation that does not provide the claimed 'non-uniform' transformation, which the parties' experts agree requires a non-linear transformation." The Board resolved this in Petitioner's favor by adopting the broad construction of "non-uniformly shifts."
3. Cross-field application (graphic-design page resizing → camera media display), which PO attacked and the Board rejected.

---

Ground III: Jeffrey (US 2009/0202177 / US 8,086,073) + a camera/display reference

Where it lands: claims 2–5, 8–10, 12–18, 20 — and this is the combination I would rank most efficient on the merits for claim 2 and claim 13, because Jeffrey's specification contains the express 4:3 → 16:9 statement.

Mapping:

• Jeffrey's resizer defines a uniformly scaled center region and non-uniformly scaled left/right regions (Fig. 4A/4B: InUni 404 of 5 pixels → OutUni 414 of 8 pixels; InNonUniL/R of 2 pixels → OutNonUniL/R of 4 pixels) → the "portion… subset of the pixels" whose displacement differs from the others (claims 1(e)/2(c) and the claim-1(h) relationship).
• Explicit purpose: 4:3 → 16:9 scaling "while minimizing distortion between the center region of the original image and the center region of the resized image."
• Implementation in a graphics controller / Mobile Graphic Engine driving a display device 212 — "displays for mobile phones, portable video game systems and smart phones."
• Non-uniform scale factor ΔS applied incrementally with error-correction values → a position-dependent mapping, satisfying the broad "non-uniformly shifts" construction.

Closing the camera gap: combine with Meulen (camera + LCD) or US 8,913,141 (GoPro's in-camera warping hardware), or simply rely on the admitted state of the art that cameras of the era had on-board image-processing pipelines and LCDs.

Motivation: Jeffrey's own specification supplies it — a person seeking to fit 4:3-captured content to a 16:9 destination wants precisely the center-preserving, edge-absorbing behavior Jeffrey describes. Substituting a camera's processor/LCD for the mobile-handheld graphics controller is the substitution of one known equivalent display pipeline element for another.

Weakness: Jeffrey's embodiment is display-pipeline-oriented and its "portion" is a fixed geometric center region, so Jeffrey is weakest on the "determine a portion" step and on claim 11/19 (object-based portion selection).

---

Ground IV: Gleicher + Oosawa (+ a "determining" reference) — the examiner's own combination

This is the rejection the applicant actually faced and only escaped by amendment.

• Gleicher (US 8,218,895) — non-uniform warped-image generation (fish-eye warping): position-dependent remapping.
• Oosawa (US 7,496,243) — "the amount of shift … is smaller in the center of the image than in the peripheral of the image": the claim-1(h)/2(f)/13 "subset moves less than others" limitation.

The applicant amended in "determin[ing] a portion within the input image" (the original claims only required that the image "includes a portion"). Note the significance: the record states the examiner's rejection was the only rejection issued during prosecution, and it was Gleicher + Oosawa. So the entire patentability margin of claims 1/2/13 rests on that added step.

To resurrect this ground, add any reference that determines the region after image acquisition:

• Philips US 2011/0262055 — saliency detector 709 → accumulator 710 → cost calculator 711 → curve generator 713, using CV-library face/eye/lip detection, and expressly handling the case where "most important object(s) are near the side panels of the screen" (this also anticipates claims 9/17, non-center portion).
• US 8,723,959 (DigitalOptics) — face/object tracking in off-center peripheral regions for nonlinear geometries (claims 9/11/17/19).
• Hwang (ICME 2008) — perceptual seam carving with a human attention model — explicitly selects where not to distort based on detected attention (claim 11/19).

Motivation: if the region of distortion is known to be a design choice (as PO's own expert conceded about Meulen — "a designer's choice," "a choice by the person who implement[s] the algorithm"), then automating that choice with a known saliency/face detector is the epitome of an obvious automation. KSR, 550 U.S. at 421; Perfect Web Techs. v. InfoUSA, 587 F.3d 1324 (Fed. Cir. 2009).

---

Ground V (claim 11/19 only): Fleming (or Meulen/Jeffrey) + DigitalOptics '959 or Hwang

Claims 11/19 require the portion to be determined based on the location of an object so the portion includes the object. That is exactly what:

• DigitalOptics US 8,723,959 does (object/face tracking to place a nonlinear-geometry region of interest), and
• Hwang's perceptual seam carving does (attention model), and
• Philips US 2011/0262055 does (saliency).

Motivation: the '413's own specification states the portion "may dynamically change between images (e.g., to follow a moving object)" and that warping "may be applied in a manner that reduces distortion applied to a particular object in motion… since the viewer is likely to focus attention on the object in motion." That is a statement of the problem the secondary references were already solving — the strongest possible motivation-to-combine evidence, because it comes from the patent itself.

---

Ground VI (claims 3–5 / 14–15 alternative): content-aware retargeting NPL

Wang et al., Optimized Scale-and-Stretch for Image Resizing (ACM TOG 2008) teaches a 2-D, non-uniform, content-aware warp that changes image aspect ratio while minimizing distortion of important content and allowing the "important" region to move — structurally the '413's core idea. Avidan & Shamir (2007) and Hwang (2008) supply the same teaching with different mechanisms. All three are of record (examiner-cited Non-Patent Citations). Motivation: the references' own stated objective is aspect-ratio change with minimized perceived distortion — the '413's stated objective verbatim.

---

4. Objective indicia (secondary considerations) and how they were treated

The Institution Decision reportedly rejected Patent Owner's evidence on objective indicia — copying and commercial success. Patent Owner's narrative (from the Preliminary Response) is that it "intentionally introduced distortion" where the prior art "sought to remove any distortion," and that this yielded an immersive "barrel" effect and user-recognized value. Two problems with that as an obviousness rebuttal:

1. The "teaching away" framing fails on the record. Distortion-avoidance ("correct the geometry") and distortion-distribution ("put it where nobody looks") are not opposite teachings. Meulen, Oosawa, and Jeffrey each deliberately concentrate distortion at the edges and preserve the center — i.e., the art was already teaching selective, non-uniform distortion. A reference does not teach away merely because it also values fidelity in the region of interest.
2. "Copying" requires nexus to the claimed subject matter, and commercial success must be tied to the claimed feature rather than to GoPro's brand, ruggedization, or distribution. The Board's rejection of this evidence in the ID suggests it was not found probative.

The strongest genuine non-obviousness argument in the record is not technical at all — it is Patent Owner's reasonable-expectation-of-success/motivation attack on the in-camera implementation: that 2013-era camera processors were resource-constrained, that "conventional wisdom in the art was to apply simpler, alternative techniques such as cropping, padding, or linear scaling," and that the prior art therefore gave no reason to move the conversion on-camera rather than leave it to the display or to post-processing. That argument worked for claim 1 and (for Fleming/Obara) for claims 6–7, and it is the argument any defendant must now overcome with evidence of contemporary on-camera image-processing capability (GoPro's own US 8,913,141 and the admitted 1080p/4K action-camera pipelines are useful for that).

---

5. Bottom line

Net assessment: On the merits, the '413's independent claims are weak. The transformation itself — a 2-D, non-uniform, position-dependent warp that preserves the region of interest and pushes distortion to the edges and corners — was squarely in the public domain by 2013 (Fleming 2007; Jeffrey 2009; Oosawa; Gleicher; Meulen; Philips; plus the seam-carving/scale-and-stretch literature). The claim that survived the IPR did so because the Board adopted a narrow reading of claim 1's scope, not because the claimed subject matter was technically distinct. Everything above is, of course, academic as an enforcement matter: the record shows statutory disclaimers under 35 U.S.C. § 253 with an effective date of 2024-11-05 covering claims 2–5, 8–20 and claims 1, 6, and 7 — i.e., the entire claim set.

---

6. Contradiction / uncertainty flags

1. ⚠️ Internal contradiction in the claim-1 rationale (flagging explicitly, as instructed). The previously-generated section and the third-party FWD summary both state that claim 1 survived because "media source content" was construed to exclude content captured by the device's internal camera. This sits in direct tension with (a) claim 1's own text, which requires "an image sensor configured to capture an input image having a source aspect ratio" — i.e., claim 1 by its terms covers internally-captured content; and (b) the Board's Institution Decision, which found "the plain meaning of the recited 'image capture apparatus' encompasses Meulen's wireless device." Note also that "media source content" is Meulen's terminology, not the '413's — the phrase does not appear in the '413 claims. I therefore treat the reported FWD rationale for claim 1 as unreliable as summarized; the actual basis should be checked against the FWD itself. What is not in doubt is that the FWD left claim 1 unpatentability unproven while striking claims 6 and 7.
2. Reference "Obara" is unidentified. It is the petition's second primary reference. I could not determine from the retrieved record which document it is. It is not obviously in the '413's citation tables. Do not treat any statement about Obara's disclosure as verified.
3. Fleming's status. US 7,158,158 B1 appears in the '413's "Family Cites Families (35)" table (not the "Citations (22)" examiner-cited table). It is nonetheless part of the supplied record and is prior art by date (2003/2007). Citing it as "of record" is accurate; calling it "examiner-cited" would not be.
4. Dependent-claim numbering. The FWD addressed only claims 1, 6, and 7; claims 3–5, 8–12, 14–20 were not adjudicated in that proceeding and my mapping for them is an analytical projection from the same references.
5. One search was cut off before returning (a query on US 7,158,158's abstract). The Fleming content used above comes from the IPR petition excerpt and the patent PDF returned by the other query; the abstract was not independently retrieved.
6. Dates. The task states the current date as April 26, 2026; the Google Patents legal-events feed and the PTAB docket material retrieved here include entries dated later in 2026 (e.g., a CAFC case-management filing and a December 2025 maintenance-fee payment). Where those conflict, I have reported the docket-entry content as it appears in the retrieved sources without reconciling the dates.