US 10767981

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

Let me dig deeper into litigation, CAFC dockets, and current ownership.

Let me check for litigation/PTAB/CAFC activity specifically involving this patent and its family, and confirm the current owner.

Summary of U.S. Patent No. 10,767,981 (B2)

Bottom line up front: I found no CAFC 2026 docket activity for this patent, and no 2026 district-court complaint asserting it. The searches for "10767981" in a litigation/CAFC context returned only unrelated 2026 matters (e.g., Taction v. Apple on 10,659,885 / 10,820,117; Wapp Tech v. Apple; Adeia v. Fubo; Adeia v. DISH). None of those involve the '981 patent. I flag this as an absence of evidence rather than proof of non-litigation, since I could not query PACER/CAFC's docket system directly.

---

Bibliographic data

Assignee note / uncertainty: The authoritative Google Patents record lists "Fotonation Ltd" as current assignee, but multiple secondary sources (patentguru, onscope, Google's patent-family listing for later continuations) show Adeia Imaging LLC as the present owner as of 2026, consistent with the Fotonation→Adeia corporate renaming/reassignment. I did not verify the chain of title in the USPTO Assignment database directly, so treat the "Adeia Imaging LLC" attribution as reported, not confirmed.

---

Abstract

The patent discloses systems and methods for estimating depth from projected texture using camera arrays. One embodiment includes at least one two-dimensional array of cameras, an illumination system configured to illuminate a scene with projected texture, a processor, and memory containing an image processing pipeline application and an illumination system controller application. The controller directs the illumination system to project texture; the pipeline captures a set of images of the illuminated scene and determines depth estimates for pixel locations in an image from a reference viewpoint using at least a subset of those images. Generating a depth estimate for a given pixel involves identifying corresponding pixels across the images based on expected disparity at multiple depths along multiple epipolar lines aligned at different angles, comparing similarity of the corresponding pixels at each depth, and selecting the depth with the highest degree of similarity.

---

Plain-language overview of the independent claims

Important caveat: The text I was given contains the full description and summary but the claims section is truncated (it cuts off in the Detailed Disclosure). The overview below is reconstructed from the patent's own "Summary of the Invention," which in this family typically tracks the claim language closely. Claim numbers and exact wording should be verified against the official USPTO claims, particularly because later continuations in the same family carry materially different claims (see "Family" note below).

Independent Claim 1 — Camera array with projected-texture depth estimation (broadest concept).
A camera array comprising:

• at least one two-dimensional array of cameras (multiple cameras);
• an illumination system that projects a texture onto the scene;
• a processor and memory storing an image-processing pipeline application and an illumination-system controller application.

Operation: the controller causes the illumination system to project texture; the pipeline captures a set of images of the textured scene and computes depth for pixels in an image from a reference viewpoint. For each pixel, depth is found by:

1. identifying corresponding pixels in the image subset based on expected disparity at several candidate depths, searched along multiple epipolar lines aligned at different angles (a consequence of the 2-D camera layout);
2. comparing similarity of the corresponding pixels at each depth; and
3. selecting the depth at which the corresponding pixels are most similar.

Independent Claim (a later "still yet another further embodiment") — Complementary-occlusion-zone configuration.
At least a pair of arrays of cameras located in complementary occlusion zones on either side of the illumination system (each array having multiple cameras), plus the illumination system, processor, and memory, performing the same different-angle epipolar-line depth estimation described above.

Selected dependent limitations (from the Summary's embodiment list, which mirrors dependent claims):

• Two 2-D arrays in complementary occlusion zones surrounding the illumination system; occlusion in one array compensated by visibility in the other.
• Each array being a 2×2 array of monochrome cameras.
• Asymmetric projected texture: a smaller spatial pattern period in one direction and a larger period in a second direction (to match anisometric baselines / a dominant epipolar direction).
• One 2-D array containing lower-resolution cameras plus at least one higher-resolution camera (e.g., higher-res Bayer camera with lower-res monochrome cameras above/below/left/right), with the higher-res camera capturing natural texture when the projector is off.
• Static vs. dynamic illumination: fixed DOE pattern vs. a controllable pattern (e.g., reflective LCOS microdisplay or transmissive/translucent LCD spatial light modulator).
• Closed-loop / adaptive texture: project a first texture, derive initial depths, then project a second texture selected from those initial depths (with the second texture's spatial pattern period finer, at that depth, than the cameras' spatial resolution there).
• Array of projectors, optionally projecting different patterns / different spatial pattern periods / controllable patterns; random patterns; and periods differing across two perpendicular directions.
• Natural-texture + depth collocation: capture images with texture projected, then without, and collocate natural texture with depth (either assuming a common viewpoint, or by estimating relative pose from features in both depth maps and reprojecting, then compositing based on depth reliability/confidence maps).

---

Family / related versions (reported, not independently verified)

• Parent: Ser. No. 14/547,048 (2014-11-18).
• Published application: US 2019/0063905 A1 (2019-02-28).
• Later continuations in the same family (per search results): US 11,486,698 B2 (2022-11-01, Fotonation Limited); US 2023/0152087 A1 (2023-05-18); US 12,590,799 B2 (reported 2026-03-31, Adeia Imaging LLC); and a 2026 publication US 2026/0168787 A1 attributed to Adeia Imaging LLC.
• Significant claim-scope difference in the continuations: the 2023 publication (US 2023/0152087 A1) recites an independent claim that adds a "conventional camera" having higher resolution than the array cameras, a horizontal baseline larger than the vertical baseline, and a final step of "process the set of images using the depth estimates to produce a synthesized higher resolution image." That combination is not reflected in the '981 abstract/summary, so do not assume the '981 claims read on the "conventional camera + super-resolution" scope — that appears to be a distinct continuation claim.

---

Prior art / references cited in the specification

• Nishihara, "PRISM: A Practical Real-Time Imaging Stereo Matcher," MIT AI Lab (A.I. Memo 780), May 1984 — projector-based binocular stereo with unstructured/random texture.
• J. Lim, "Optimized projection pattern supplementing stereo systems," ICRA 2009 — De Bruijn-sequence patterns.
• K. Konolige (spelled "K. Klonige" in the text), "Projected Texture Stereo," ICRA 2010 — Hamming-code-based patterns.
• Incorporated-by-reference patents: US 8,619,082 (Ciurea et al., depth/parallax and confidence maps); US 8,780,113 (Ciurea et al., depth using multiple spectral channels); US 2013/0293760 (Nisenzon et al., π filter groups); US 2011/0069189 (Venkataraman et al., heterogeneous imagers); US 2012/0012748 (Pain et al., SoC array cameras).

---

Uncertainty disclosures

1. Claims text: I could not reproduce the verbatim granted claims of US 10,767,981 from the material provided; the overview above is inferred from the Summary of the Invention and should be confirmed against the official claim set.
2. Litigation/CAFC: No 2026 CAFC docket or district-court assertion of US 10,767,981 surfaced. I could not run a direct PACER/CAFC docket query, so this is a negative search result, not a guarantee.
3. Ownership/status dates: "Active," "adjusted expiration 2034-12-16," and the 2022/2026 continuation issuance dates come from the Google Patents page and search snippets; some 2026-dated items in the family are unusually recent and were not verified at source.

I'll search for litigation involving US patent 10767981 using the sources you specified.

Let me search more specifically on litigation databases.

Bottom line

Based on the searches I was able to run, I found no litigation (district court, ITC, PTAB/IPR, or UPC) specifically identifying U.S. Patent No. 10,767,981 ("the '981 patent") as an asserted or challenged patent. I therefore cannot populate a case table with plaintiff/defendant/case number/date data for this patent.

I want to be explicit about my confidence level: this is a negative finding from a limited search, not a verified "no litigation exists" conclusion. Treat it as "not found," not "confirmed none."

What the searches did surface

1. The patent itself — Unified Patents portal

• URL: https://portal.unifiedpatents.com/patents/patent/US-10767981-B2
• The Unified Patents page exists for US-10767981-B2 and shows the patent's bibliographic data and cited "Patent Art," but the snippet returned no litigation section and no listed cases.
• Note: Unified's litigation data is populated for patents that have been asserted in tracked cases. The absence of a litigation listing there is consistent with, but not proof of, no litigation.

2. The other results were false positives — different patent numbers. Per your strict rule, I did not treat these as matches to 10,767,981:

• Jawbone Innovations LLC cases (Amazon, Google, etc.) concern a patent referred to as the "'091 patent" — different number.
• Headwater Research v. Motorola/Lenovo (N.D. Cal. 4:23-cv-04496) concerned U.S. 9,198,076 and 10,749,700 — different numbers.
• Taction Technology v. Apple concerned 10,659,885 and 10,820,117 — different numbers.
• Other results involved unrelated patents (e.g., '302, '851, '097, '700 patents).
• Several results surfaced unrelated patents that merely cite or are cited by the '981 patent in citation lists (e.g., US 10,366,472; US 10,542,208; US 10,542,113 family references). Citation relationships are not litigation.

3. Ownership context relevant to any future assertion (from the patent record, not from litigation filings)

• Original assignee: FotoNation Ireland Ltd; current assignee listed as FotoNation Ltd (Google Patents page).
• A 2023 assignment of security interest to Bank of America, N.A. as collateral agent, with assignors including Adeia Imaging LLC and other Adeia entities — indicating the patent now sits in the Adeia (formerly Xperi/Tessera) portfolio. This is an ownership/security record, not litigation.
• Continuation/priority chain: continuation of App. Ser. No. 14/547,048 (filed Nov. 18, 2014), claiming priority to Provisional 61/905,423 (Nov. 18, 2013). Related granted family members noted on the record include US 11,486,698 and US 12,590,799 — again, family relationships, not litigation.

Suggested verification steps (since my search was not exhaustive)

To close this out with confidence, I'd check these directly rather than relying on general web search:

1. Unified Patents litigation portal — https://portal.unifiedpatents.com/litigation/caselist — search the patent number in the "Asserted Patents" filter.
2. RPX Insight — https://insight.rpxcorp.com/ — search "10,767,981".
3. CourtListener / RECAP — https://www.courtlistener.com/ — full-text search "10,767,981" and "10767981" across dockets.
4. USPTO PTAB — https://ptacts.uspto.gov/ (and the PTAB "Patent Trial" search) — search the patent number for any IPR/PGR/CBM and check the "Related Matters" section of any petition naming it.
5. USPTO Patent Assignment Search — https://assignment.uspto.gov/ — to confirm any transfer that typically precedes assertion.
6. PACER — confirm any hit found in steps 1–4 at the district level.

Direct answer to your request

• Plaintiff(s): None identified.
• Defendant(s): None identified.
• Jurisdiction: N/A.
• Case number: N/A.
• Filing date: N/A.
• Outcome/current status: No litigation known to me for US 10,767,981 based on the searches performed.

If you want, I can run the targeted queries against CourtListener, PTAB, and RPX specifically and report back case-by-case — but I did not want to present any of the above false-positive results as if they were litigation involving the '981 patent.

I'll start by checking the structured data and then searching for any PTAB proceedings on this patent.

Let me dig deeper into patent-specific PTAB records and any related proceedings.

Let me verify directly against PTAB/court databases and check the patent's litigation/assignment history.

PTAB Proceedings on US 10,767,981 — Analyst Report

Data currency: Structured "PTAB proceedings on file" block (USPTO Open Data Portal, most recent ingest) + independent web checks, 2026-09-29.

---

Proceedings overview

Total AIA trial proceedings on file: 0 — no Inter Partes Reviews, no Post-Grant Reviews, and no Covered Business Method reviews have ever been filed against US 10,767,981, so the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials; the patent is entirely untested at the PTAB, which means it has not been narrowed by any Final Written Decision, all of its issued claims remain live and presumptively valid under 35 U.S.C. § 282, and a defendant today gets no free "claim 1 is canceled" argument — an IPR-based defense must be built from scratch (or the case must be fought on district-court invalidity/§ 282, ex parte reexam, or non-infringement).

Important caveat on scope of this report: Because the canonical proceedings list is empty, there are no {PROCEEDING_NUMBER} — {Petitioner} v. {Patent Owner} entries to populate below. I am not inventing proceeding numbers. What follows is (a) an explicit record of the negative finding and how I checked it, (b) one unverified web-search artifact that must be run down before you rely on the negative, and (c) the strategic read on what an empty PTAB docket means for a defendant.

Verified negative finding — no proceeding block to reproduce

• Type: n/a — no AIA trial petition of any type has been accorded a filing date against this patent.
• Filed: n/a
• Status: The structured ODP "PTAB proceedings on file" block returns an empty set ("no AIA trial proceedings… as of the most recent ingest"). Independent web checks (Google Patents legal-events/assignments page, Unified Patents patent portal entry for US-10767981-B2, and PTAB document searches) surfaced no IPR/PGR/CBM bearing this patent number, no Final Written Decision, and no Federal Circuit appeal arising from a PTAB proceeding on it. Treat this as "no PTAB activity on file."
• Judge panel: n/a
• Petition grounds: n/a
• Institution decision: n/a
• Final Written Decision: n/a — no FWD exists, so no claim of US 10,767,981 has ever been canceled or confirmed by the Board.
• Settlement / termination: n/a
• Appeal: n/a — no PTAB appeal to the Federal Circuit exists for this patent (there is likewise no CAFC docket on CourtListener tied to a PTAB decision on US 10,767,981).
• Defensive value: Neutral-to-negative for a defendant. There is no Board ruling, claim cancellation, or estoppel to leverage, and no PTAB-tested claim construction to borrow. You also inherit a clean slate: no § 315(e)(2) estoppel has been triggered by anyone against this patent, so your prior-art runway is wide open.

⚠️ Unverified artifact to run down before you finalize the negative

Web search surfaced a Patent Owner Preliminary Response (POPR) excerpt in which the patent owner argues General Plastic against a follow-on petition: the petitioner filed IPR2021-00881 on 2021-05-05 against a "'338 Patent" and roughly fourteen months later filed a new petition against a related "'681 Patent" sharing the same specification, asserting the '681 claims "mimic" (petitioner argued "identical") the '338 claims, using the same references (including an "Exclusion Criteria" limitation discussion).

I could not confirm that this "'681 Patent" is US 10,767,981. The surrounding subject matter ("Exclusion Criteria," a '338 sibling patent) does not obviously match the projected-texture/camera-array family of US 10,767,981 (whose continuations are US 10,119,808, US 11,486,698, and US 12,590,799). It is at least as likely a different family entirely. Do not treat this as a proceeding on 10,767,981. Verify by pulling the POPR at PTAB E2E (https://ptacts.uspto.gov/) and checking the patent number on its cover. If it turns out to be this patent, this section changes materially — an IPR2021-00881-adjacent petition would be a proceeding on file and the ODP negative above would need correction.

---

Strategic summary

Claim status: everything is UNTESTED; nothing is canceled, nothing is confirmed. No independent claim and no dependent claim of US 10,767,981 has been through an AIA trial. That is a two-edged fact. On the favorable side for the patent owner, they carry zero adverse claim-cancellation precedent and zero Board-authored claim constructions narrowing their infringement theories. On the unfavorable side for them, they also have no res judicata-backed validity ruling and no Board "sustained" claims to brandish — the patent's validity is simply unresolved at the agency level. Practically, for a defendant, "0 proceedings" means there is no FWD you can point to and no canceled-claim carve-out; the entire claim set (whatever the cover sheet shows — verify the issued claims on the Google Patents "Claims" tab, I did not independently confirm the claim count here) is still in play. The patent's term runs to an adjusted expiration of 2034-12-16 with priority to 2013-11-18, so a long enforcement tail remains, and the family is still being prosecuted (recent issuance US 12,590,799 listed with a 2026-03-31 date on Google Patents), meaning the owner can keep minting continuations with retooled claims.

Estoppel landscape: no § 315(e)(2) estoppel is running against anyone. Because no IPR/PGR was ever instituted, no petitioner (or its privies/real parties in interest) is estopped in a parallel district court action from raising any § 102/§ 103 ground. Conversely, you are not constrained by anyone else's estoppel either — there is no prior Board record to inherit. That is the single most defendant-friendly consequence of the empty docket. It also means if you file first, you capture the benefits of being the petitioner (control of grounds, the § 315(b) one-year timeline running from your own service date), and you drive the Board's construction before the district court reaches trial. Watch, however, that the patent owner has not been through a POPR on this patent, so you have no preview of their obviousness-rebuttal themes (secondary considerations, teaching-away, priority-date attacks) — you are pleading in the dark on their strongest arguments.

Pattern signals: no petitioner, no defensive aggregator, no attack pattern — and that silence is itself informative. A search of the Unified Patents portal entry for US-10767981-B2 shows the usual prior-art landscape but no Unified-filed IPR against it (Unified's IPR program is visible elsewhere in the record, e.g., Unified Patents, Inc. v. Personalized Media Communications, IPR2015-00521, and Unified Patents, LLC v. Bradium Technologies, IPR2018-00952 — neither concerns this patent). Notably, this patent belongs to the Fotonation/Pelican array-camera portfolio that has been actively monetized and is now held in the Adeia Imaging / Adeia Inc. corporate family: the chain of title runs Fotonation Cayman → Fotonation Limited (Ireland) → Adeia Imaging LLC, and on 2023-05-03 a security interest was recorded in favor of Bank of America, N.A. as collateral agent against a roster of Adeia entities — i.e., the portfolio is collateralized and the owner has a standing licensing/assertion apparatus. That a collateralized, actively-licensed portfolio whose patents sit in heavy multi-camera phone/AR/automotive technology has attracted few or no PTAB challenges to the '981 specifically suggests either (i) it has not been put in front of a well-resourced defendant in a way that triggered a validity fight (settlements/licensees choosing not to challenge), or (ii) challenges were considered and declined on cost/merits grounds. Either way, the "well-asserted patents eventually attract IPRs" heuristic has not fired here — which cuts against an assumption that the claims are easy to knock out.

---

Recommended next steps

• Do not wait for an FWD that does not exist. There is no decision to cite, no canceled claim to quote, and no PTAB E2E link to an opinion on US 10,767,981 to hand you. Any privilege communication or demand-response letter asserting that "the claims have been invalidated" would be false as of 2026-09-29.
• Confirm the negative at the source before relying on it. Pull the patent's litigation/proceeding history at PTAB E2E (https://ptacts.uspto.gov/) and the assignment/legal-events tab on Google Patents (https://patents.google.com/patent/US10767981/en). Specifically, run down the IPR2021-00881 / "'681 Patent" POPR artifact flagged above — if that '681 is this patent, the entire "0 proceedings" conclusion must be revised.
• Weigh filing your own IPR, mindful of the § 315(b) clock. If you have been served with a complaint asserting this patent, § 315(b) bars a petition more than one year after service — calendar that date immediately. If you are not yet in suit, an IPR (or, for pre-AIA-style art/printed-publication issues, an ex parte reexamination under § 302) is the cleanest path to a Board-authored construction and a certificate canceling claims, with the added benefit of a possible district-court stay.
• Prior-art runway is unrestricted — use it. With no § 315(e)(2) estoppel and no prior Board record, you may assert any § 102/§ 103 ground, including art any earlier hypothetical challenger "could have raised." This is the rare posture where a well-built obviousness case is not hemmed in by someone else's failed petition.
• Stress-test the priority date. Priority is claimed to 2013-11-18 (provisional 61/905,423) via continuation of 14/547,048; the family chain runs through a 2014-11-18 non-provisional/PCT (WO 2015/074078). Art dated 2013–2015 must be scrutinized against swear-behind/priority challenges; the specification is dense (illumination systems, DOEs, SLMs, projector arrays), so § 112 written-description/enablement and § 101 attacks on the computer-implemented depth-estimation claims are also worth an independent look outside the IPR track — those are exactly the grounds an IPR cannot reach (§ 311(b) limits IPR to § 102/§ 103).

Sources used: Google Patents US10767981B2 (https://patents.google.com/patent/US10767981/en); Unified Patents portal (https://portal.unifiedpatents.com/patents/patent/US-10767981-B2); PTAB E2E (https://ptacts.uspto.gov/); Justia patent page (https://patents.justia.com/patent/10767981). No PTAB Final Written Decision or Federal Circuit opinion on this patent was located to cite.

I'll research the assignment record for US 10,767,981 across the sources you specified.

Let me dig into the ownership chain and specific assignment records.

I found the key reel/frame references. Let me dig deeper into the chain of title.

I've hit my search budget for this session. Let me report what the available records establish, and be explicit about the gaps.

---

Ownership Chain — US 10,767,981 B2

Sourcing note / limitation. I was not able to complete a direct pull of the USPTO Assignment Center record for the '981 patent within this session (the Assignment Center URL is https://assignmentcenter.uspto.gov/). The timeline below is reconstructed from the patent's own front-matter and Google Patents legal-events record (authoritative here), cross-checked against SEC filings and PTAB chain-of-title filings that name the Adeia/FotoNation recording reels. Where I could not retrieve a reel/frame or a correspondent, I say so rather than infer one. Treat unverified reel/frame citations as leads to confirm in Assignment Center, not as established records.

Inventors

Unusual pattern flag. The inventors' employer (Pelican Imaging Corporation, Santa Clara, CA) is not the assignee named on the issued patent (FotoNation Ireland Ltd / FotoNation Ltd). That mismatch is the tell that the asset passed through at least one intervening acquisition/assignment between invention and issuance — which is exactly what the corporate chain below shows. I found no evidence of inventors departing within 12 months of filing; both were Pelican personnel across the 2013–2014 array-camera filings.

Original assignee

• Entity named on the granted patent: the Google Patents record lists the Original Assignee as FotoNation Ireland Ltd, with the current assignee as FotoNation Ltd (Ireland). Both are Irish entities in the Tessera → Xperi → Adeia lineage.
• Entity that actually made the invention: Pelican Imaging Corporation (Santa Clara, CA), applicant on the parent application Ser. No. 14/547,048 (filed 2014-11-18).
• Business / product status: Pelican Imaging was a fabless array-camera / light-field imaging developer that licensed its technology; it did not achieve mass-market hardware shipments embodying these claims. It ceased to be independent when the Tessera/FotoNation group acquired its assets (~2016), after which the portfolio moved under FotoNation Cayman Limited → FotoNation Limited → Xperi → Adeia.
• Current status: record owner FotoNation Ltd, a subsidiary in the Adeia (formerly Tessera/Xperi) IP-licensing family. Adeia Inc. is a public company; per its SEC filings its sole reportable segment is "Intellectual Property (IP) Licensing." It does not manufacture products embodying the claims.

Assignment timeline

The package's own legal-events record exposes only one post-issuance assignment (the 2023 security interest). The ownership transfers that produced the mismatch between inventor employer and record assignee are reflected as corporate acquisitions/reorganizations rather than as a clean sequence of reel/frame records I could verify. Reported as follows:

• (2013-11-18) — Provisional 61/905,423 filed by Pelican Imaging (priority anchor). Not an assignment.
• (2014-11-18) — Parent application Ser. No. 14/547,048 filed; applicant Pelican Imaging Corporation. Not an assignment.
• (≈2016, executed/recorded dates unverified) — Acquisition / asset transfer: Pelican Imaging portfolio → FotoNation (Tessera) family. Reel/frame not verified. Reel 034745/0158 appears as a chain-of-title link in a later Adeia PTAB filing but I could not confirm it covers the '981 patent. Context: acquisition (startup absorbed into a licensing group).
• (2018-10-31) — Continuation Ser. No. 16/177,191 filed by FotoNation Ireland Ltd (the application that matured into the '981 patent).
• (2020-09-08) — Patent granted; assignee of record FotoNation Ltd.
• (2022-10-01) — Corporate separation: Xperi Holding Corporation split into Xperi Inc. (products) and Adeia Inc. (IP licensing); the licensing portfolio, including this family, went to Adeia. Context: internal reorganization / corporate separation.
• (2023-05-03) — Conveyance: Security Interest (Security Agreement) to BANK OF AMERICA, N.A., as Collateral Agent. Assignors listed include ADEIA IMAGING LLC, ADEIA MEDIA HOLDINGS LLC, ADEIA GUIDES INC., ADEIA MEDIA SOLUTIONS INC., and the Adeia Semiconductor entities. Reel/frame not retrieved directly; a related Adeia chain-of-title filing cites Reel 067609/0435 (addressed to Adeia Media Holdings LLC), which is consistent with a 2023 recording but is unverified for the '981 patent. Context: securitization (collateral pledge; does not convey right/title/interest).
  • Correspondent: not retrieved — flagged as a gap. If pulled, the recurrence test (same recording attorney across the Adeia/Marathon/etc. chains) should be applied here.

No post-issuance assignment to an anonymous shell LLC and no pre-suit transfer appear in the retrievable record.

Timeline diagram

timeline
    title Ownership of US 10767981
    2013 : Provisional filed by Pelican Imaging
    2014 : Parent application filed
    2016 : Pelican Imaging acquired by FotoNation
    2018 : Continuation filed by FotoNation Ireland
    2020 : Patent granted to FotoNation Ltd
    2022 : Xperi separation creates Adeia
    2023 : Security interest to Bank of America

NPE / troll-pattern signals

1. Shell-entity transfer — unclear. The chain terminates in LLCs (Adeia Imaging LLC, Adeia Media Holdings LLC) that sell no products, but these are subsidiaries of a public registrant (Adeia Inc.), not anonymous single-purpose vehicles. I could not verify registered-agent addresses or single-member status. No shell finding on the evidence available.
2. Known asserter in the chain — unclear, leaning present. The current owner (Adeia, and predecessors Tessera/Xperi) is a high-volume patent licensor/litigant commonly tracked by RPX/Unified. I could not retrieve a specific RPX/Unified listing in this session, so I do not assert this as established.
3. Repeat correspondent across the chain — unclear / not established. The correspondent of record was not retrievable here, so the recurrence test cannot be run. This is the single most important missing data point.
4. Cascading transfers <24 months — not present. The chain is a ~9-year sequence of an acquisition (≈2016) + corporate reorganizations (2022) + a security pledge (2023), not rapid back-to-back LLC hops.
5. Pre-litigation transfer — not present. No litigation naming the '981 patent was found (per the litigation section), so there is no pre-suit assignment to test.
6. Bankruptcy fire-sale — not present / unclear. Pelican Imaging was absorbed via acquisition; I found no Chapter 7/11 sale record for this asset.
7. Privateering — unclear. The portfolio pattern (startup's patents aggregated into a licensing company) is aggregator-like, but I have no SEC/EFF/Patent Progress evidence of an operating-company principal asserting on its behalf.
8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT, Unified, or OIN; it terminates in a for-profit licensing group.

Verdict

NPE — moderate confidence.

The chain ends at a publicly documented non-practicing IP-licensing owner: the record assignees are FotoNation Ltd and Adeia holding LLCs (Adeia Imaging LLC / Adeia Media Holdings LLC), and Adeia Inc. reports its only business segment as IP Licensing, with no products embodying the claims. That "transfer into a licensing-only owner, originating from an acquired startup portfolio" is the one moderate-quality signal supporting an NPE classification. Confidence is held at moderate, not high, because the classic troll markers are absent or unverified: there is no anonymizing shell-LLC cascade, no repeated NPE recording attorney (correspondent not retrieved), and no litigation of record for this patent — so this reads as a large, transparent, publicly traded licensing company rather than a shell-troll chain.

Verify at: https://assignmentcenter.uspto.gov/ — search patent number 10,767,981; then pull the correspondents on the Pelican→FotoNation transfer and on the 2023 Bank of America security-interest recording to complete signals 2 and 3.

Data gaps that could flip this call: (a) the correspondent attorney on each recording (signal 3); (b) confirmation of Reels 034745/0158 and 067609/0435 as covering the '981 patent; (c) any RPX/Unified assertion listing for Adeia/Xperi/Tessera naming this patent.

I'll research US 10,767,981's citation record and prior art. Let me run targeted searches.

Let me pull the "References Cited" front-page data for the '981 patent.

Let me try to retrieve the patent's front-page citation list directly.

Prior Art Analysis — U.S. Patent No. 10,767,981 (US10767981B2)

1. Patent identification (confirmed)

I verified the number literally and did not substitute any near-neighbor (e.g., US 10,366,472; US 10,542,208; US 10,749,700; US 10,659,885) that appear in unrelated search hits or in citation lists.

---

2. Methodology and an important limitation (read this before using the table)

Two categories of "citations" attach to a U.S. patent:

• (A) References cited by the applicant in the specification (the "Background" discussion and the incorporation-by-reference sentences). These are text-verifiable and I can reproduce them from the granted document.
• (B) References cited by the Examiner on the front page under "(56) References Cited" (U.S. Patent Documents, Foreign Patent Documents, Other Publications).

Limitation: Within this session I could retrieve the granted specification text and bibliographic record (Google Patents, Justia, Espacenet, FPO, Unified Patents), but I was not able to pull the complete front-page (56) Examiner-cited list for US 10,767,981 / US 2019/0063905 A1. Search results returned citation tables for other patents (e.g., US 11,053,481; US 10,119,661), which are not this patent and are excluded per the strict rule. The table below therefore reproduces the references of record in the patent text itself, which is authoritative for those items, and I explicitly flag the front-page list as not verified.

The parent PCT publication of the same application, WO 2015/074078 A1 (https://patents.google.com/patent/WO2015074078A1/en), carries the international search report that would list Examiner-type art for this family; I was not able to open that search report. That is the single most important item to retrieve next.

---

3. References cited in the '981 patent (Category A — applicant-cited of record)

Each entry: full citation · date · description · claim exposure.

3.1 Non-Patent Literature

(1) H. K. Nishihara, "PRISM: A Practical Real-Time Imaging Stereo Matcher," MIT Artificial Intelligence Laboratory, A.I. Memo 780, May 1984.

• Date: May 1984 — far predates the 2013-11-18 priority date.
• Description: Discloses binocular stereo depth estimation in which a scene is illuminated with an unstructured (random) projected texture to create matching targets where surface contrast is low; the matcher assumes no a priori knowledge of the pattern. The '981 patent's own Background § acknowledges this as the origin of "projected texture stereo."
• Claim exposure (§102): Discloses the generic concept of projecting a texture and estimating depth by disparity — i.e., the "illumination system configured to illuminate a scene with a projected texture" element. It does not disclose a two-dimensional array of cameras or correspondence search along epipolar lines aligned at different angles. Therefore PRISM alone does not anticipate independent claims 1 or 22, but it is squarely §102 anticipatory art against any hypothetical claim drawn to projected-texture stereo depth estimation per se, and is core §103 art against the projector-plus-disparity combination.

(2) J. Lim, "Optimized projection pattern supplementing stereo systems," ICRA 2009.

• Date: 2009 — predates priority date.
• Description: Proposes projected patterns generated from De Bruijn sequences to avoid self-similar regions that cause false disparity matches.
• Claim exposure (§102): Directly relevant to the pattern-selection limitations — the dependent claims reciting a random pattern and a pattern having different spatial pattern periods in orthogonal directions. Anticipates only pattern-specific dependent claims, not the array architecture of claim 1.

(3) K. "Klonige," "Projected Texture Stereo," ICRA 2010. (spelling reproduced literally from the '981 patent text; standard literature renders this author as K. Konolige — flagged, not corrected in reliance)

• Date: 2010 — predates priority date.
• Description: Hamming-code–based projection patterns to avoid self-similarity in projected-texture stereo.
• Claim exposure (§102): Same role as (2) — art against the projected-pattern characterization and the anti-self-similarity rationale, not against the 2D-array architecture.

3.2 U.S. Patent Documents (applicant-cited / incorporated by reference)

(4) U.S. Pat. No. 8,619,082 B2 — Ciurea et al., "Systems and Methods for Parallax Detection and Correction in Images Captured Using Array Cameras that Contain Occlusions using Subsets of Images to Perform Depth Estimation."

• Date: Issued 2013-12-31 (family priority pre-dates 2013-11-18). Prior art as of the '981 priority date.
• Description: Depth estimation in array cameras by disparity search using subsets of images selected to avoid occlusions, plus confidence metrics for depth reliability. The '981 patent expressly incorporates it for "depth estimation using sets of images" and for "confidence metrics."
• Claim exposure (§102/§103): Highly material to claim 1's depth-estimation core and to the confidence-map-based compositing (dependent claims). Alone it lacks projected texture, so it does not anticipate claim 1; it is the strongest §103 primary reference to combine with a texture projector.

(5) U.S. Pat. No. 8,780,113 B2 — Ciurea et al., "Systems and Methods for Performing Depth Estimation using Image Data from Multiple Spectral Channels."

• Date: Issued 2014-07-22 (priority earlier). Prior art.
• Description: Depth estimation by using corresponding pixels across multiple spectral channels to increase the number of epipolar lines searched.
• Claim exposure (§102): Directly material to the dependent claims reciting multi-spectral / π-filter-group camera arrays and to the FIG. 2B discussion of increasing epipolar-line count via color channels.

(6) U.S. Patent Publication No. U.S. 2011/0069189 A1 — Venkataraman et al., "Capturing and Processing of Images Using Monolithic Camera Array with Heterogeneous Imagers."

• Date: Published 2011-03-24. Prior art.
• Description: Monolithic camera arrays with heterogeneous imagers (different lenses/resolutions).
• Claim exposure (§102): Directly material to the dependent claims reciting a 2D array having lower-resolution cameras and at least one higher-resolution camera, with lower-res cameras above/below/left/right of the higher-res camera.

(7) U.S. Patent Publication No. U.S. 2013/0293760 A1 — Nisenzon et al., "Camera Modules Patterned with π Filter Groups."

• Date: Published 2013-11-07. Prior art (before 2013-11-18 priority).
• Description: π-filter-group camera arrangements (central Green, Red/Blue pairs in complementary zones, Green cameras).
• Claim exposure (§102): Material to the dependent claims/embodiments describing FIG. 1D two 3×3 π-filter-group arrays in complementary occlusion zones on either side of the illumination system.

(8) U.S. Pat. No. 8,878,950 B2 — Lelescu et al., "Systems and Methods for Synthesizing High Resolution Images Using Super-Resolution Processes."

• Date: Issued 2014-11-04 (priority earlier). Prior art.
• Description: Super-resolution synthesis from array-camera image data.
• Claim exposure: Supports the texture-synthesis / super-resolution step (FIG. 6, block 614). Generally §103 art, not anticipatory of any independent claim.

(9) U.S. Patent Publication No. U.S. 2012/0012748 A1 — Pain et al., "Architectures for System on Chip Array Cameras." (note: FPO rendering of the same sentence shows "2012/0013748"; literal discrepancy flagged)

• Date: Published 2012-01-19. Prior art.
• Description: Multi-focal-plane sensor / SoC array-camera architectures.
• Claim exposure: Background/structural art for how the camera array is built; not anticipatory of the depth claims.

(10) U.S. patent application Ser. No. 14/536,537 — Rodda et al., "Methods of Manufacturing Array Camera Modules Incorporating Independently Aligned Lens Stacks," filed 2014-11-07.

• Date: Filed 2014-11-07 — after the 2013-11-18 priority date. NOT prior art to the '981 patent; cited only as incorporation-by-reference for manufacturing.
• Claim exposure: None (post-priority; manufacturing-only).

---

4. Anticipation assessment — bottom line

No single reference of record appears to anticipate independent claims 1 or 22.

The independent claims require the combination of (i) at least one two-dimensional array of cameras, (ii) an illumination system projecting texture on the scene, and (iii) correspondence/“similarity” search at multiple depths along a plurality of epipolar lines aligned at different angles. The cited art splits cleanly:

• Projected-texture + disparity is old: Nishihara/PRISM (1984), Lim (2009), "Klonige" (2010) — but all are binocular/1-D baseline, not 2-D arrays with multi-angle epipolar search.
• 2-D array depth estimation with multi-angle epipolar search + occlusion tactics is old: U.S. 8,619,082 (Ciurea) — but without projected texture.

Accordingly:

• §102 anticipation is not met by any identified reference for claims 1/22. The NPL references could anticipate only narrow, pattern-focused dependent claims (e.g., random/De Bruijn/Hamming pattern or differential spatial-period claims), and U.S. 2011/0069189 could anticipate only heterogeneous-resolution dependent claims.
• The realistic invalidity theory is §103: Ciurea '082 (2-D array depth estimation via subsets/epipolar search, with confidence maps) as primary, combined with Nishihara/PRISM or Lim (projected texture to supply texture-lacking regions), and with U.S. 2013/0293760 or U.S. 2011/0069189 for the π-group / heterogeneous-resolution dependent claims. The '981 patent's own Background effectively concedes the projector-plus-stereo combination as known, which strengthens such a combination argument.

5. What I could not verify (be explicit)

1. The complete front-page (56) "References Cited" list (Examiner-cited U.S. and foreign patent documents and "Other Publications") for US 10,767,981 — not retrieved. The entries above are the applicant-cited/in-specification references, which are of record but may be a subset of the full citation set.
2. The international search report / citations for WO 2015/074078 A1, the family member whose search report would carry the Examiner-type art. Recommended to retrieve next.
3. Any §102 art cited during prosecution but not carried onto the '981 front page (e.g., art cited only in the parent 14/547,048 or in U.S. 2019/0063905 A1's record).

Recommended next sources to close these gaps: USPTO Patent Center / Patent Application Information Retrieval for App. 16/177,191 (full IFW "References Cited"), Espacenet "Cited documents" for US10767981B2 (https://worldwide.espacenet.com/patent/search?q=US10767981), and the WO 2015/074078 A1 search report on PATENTSCOPE. I did not invent any entries for these unretrieved lists.

I'll verify the actual claim language and the prior-art references before building the §103 analysis.

Let me pull the exact granted claim language and confirm the Lim/Konolige prior-art references.

Obviousness Analysis — U.S. Patent No. 10,767,981 (Venkataraman & Duparré)

Statutory basis: 35 U.S.C. § 103, as amended by the AIA (the '981 patent's claim chain has an earliest effective filing date of Nov. 18, 2013 via Provisional 61/905,423; parent Ser. No. 14/547,048 filed Nov. 18, 2014 — both post-March 16, 2013, so AIA § 102/§ 103 governs).

---

0. Scope, assumptions, and a flagged contradiction

Claim-text caveat (carried forward from the summary section): I still could not retrieve verbatim granted-claim text. The Google Patents/Justia snippets confirm that the claim 1 language tracks the "Summary of the Invention" nearly word-for-word, and I verified a separate point below that materially strengthens the analysis. The element-by-element mapping is built on that Summary language plus the independent "complementary occlusion zones" embodiment. Confirm against the official USPTO claim set before filing anything.

Important observation found during this pass — the claim language is lifted from the assignee's own earlier patent. The RPX claim record for US 8,619,082 B1 (Ciurea et al.) recites, in relevant part:

"determining initial depth estimates for pixel locations in an image from the reference viewpoint using at least a subset of the set of images, where an initial depth estimate for a given pixel location in the image from the reference viewpoint is determined by; identifying pixels … based upon expected disparity at a plurality of depths; comparing the similarity of the corresponding pixels identified at each of the plurality of depths; and selecting the depth from the plurality of depths at which the identified corresponding pixels have the highest degree of similarity…"

Compare the '981 claim 1 (per Summary): "identifying pixels … based upon expected disparity at a plurality of depths along a plurality of epipolar lines aligned at different angles; comparing the similarity of the corresponding pixels identified at each of the plurality of depths; and selecting the depth from the plurality of depths at which the identified corresponding pixels have the highest degree of similarity."

The only additions in '981 over '082's claim language are (i) the word "epipolar lines aligned at different angles," and (ii) the presence of the illumination system. That is a very small delta for a § 103 case, and the "different angles" phrase is arguably inherent in the 2D array that '082 already claims (discussed in §4).

Flagged contradiction / caveat on reference availability: US 8,619,082 issued Dec. 31, 2013, which is after the '981's Nov. 18, 2013 priority date — so it is not § 102(a)(1) art via its own issue. It could only qualify under § 102(a)(2) (effectively filed Aug. 21, 2012, naming different inventors), but if '082 and the '981 application were commonly owned (both trace to Pelican Imaging Corp. at the relevant time — '981 names Kartik Venkataraman, who is also a named inventor on '082), the § 102(b)(2)(C) common-ownership exception likely removes '082 as § 102/§ 103 art. This is a genuine weakness in any IPR built on '082 as the primary reference and must be verified in the USPTO Assignment database. I flag it rather than paper over it. The analysis below therefore also supplies § 102(a)(1)-clean primary references (Venkataraman US 2011/0069189; Pain US 2012/0012748; and the non-patent literature), so the case does not depend on '082 alone.

---

1. The person of ordinary skill in the art (PHOSITA)

A bachelor's degree in electrical engineering, computer engineering, computer science, physics, or optics, plus 2–3 years of experience in computational imaging, stereo/multi-view depth estimation, structured light, or camera-array design (or a master's degree with commensurately less experience). This person would be familiar with: epipolar geometry and disparity search; block-matching/SAD cost functions; structured-light pattern codification (Gray code, De Bruijn, Hamming); camera-array architectures; and projection optics. All of the § 103 reasoning below is pitched at this level.

---

2. The prior art of record on this page

Additional art a PHOSITA would know (outside the cited set, flagged as such): Okutomi & Kanade, "A Multiple-Baseline Stereo" (1993); Wilburn et al., "High Performance Imaging Using Large Camera Arrays" (SIGGRAPH 2005); Salvi et al., "Pattern codification strategies in structured light systems," Pattern Recognition 37(4) (2004); Scharstein & Szeliski stereo taxonomy (2002); and consumer IR-pattern-projection depth sensors (Kinect/PrimeSense).

---

3. Claim 1 — element mapping

No single reference discloses all of (a)–(j). Claim 1 is therefore not anticipated, but it is a textbook combination case.

---

4. The pivotal limitation — "epipolar lines aligned at different angles" — is inherent in the 2D array

This is the limitation the patent appears to treat as its point of novelty, but the '981 specification itself admits it is a characteristic of the pre-existing 2D camera array:

"Due to the spatial relationship of cameras in a two-dimensional array of cameras, the epipolar lines searched during the disparity search are aligned at different angles."

and

"estimating depth using a set of images captured by a two dimensional array of cameras typically involves performing disparity searches along epipolar lines aligned at different angles."

A reference that teaches depth estimation from a 2D array of cameras (Ciurea '082; Venkataraman '189; Pain '148) therefore inherently discloses or renders obvious correspondence searches along non-parallel epipolar lines, because three or more non-collinear camera centers define epipolar lines that are not parallel. The applicant's own words supply the inherency/obviousness rationale. This is the single most important point of attack.

---

5. Primary § 103 combination

Ground 1 — Ciurea '082 (primary) in view of Nishihara, further in view of Lim or Konolige

Combination: Ciurea '082 (2D array camera; reference-viewpoint depth via disparity search over multiple depths; similarity comparison; best-match selection; confidence maps; occlusion handling) + Nishihara (project a random/“unstructured” texture onto the scene to supply matching targets for stereo matching in textureless regions) + optionally Lim or Konolige (optimized non-self-similar projected patterns).

Why every element is met: As the table in §3 shows, Ciurea '082 supplies (a), (c), (d), (f), (g), (h–plurality of depths), (i), (j) essentially verbatim; Nishihara supplies (b) and (e) and the textureless-region rationale; Lim/Konolige supply the pattern-design refinements.

Motivation to combine — express in the references themselves (strongest form):

1. Nishihara states the motivation verbatim: the projector "provide[s] fixed surface markings on objects lacking sufficient natural texture," i.e., textureless regions — exactly the failure mode of a passive camera array that the '981 background concedes ("depth estimation can fail in regions of a scene that lack texture"). One reference supplying the precise motivation for combining with the other is the classic In re Kemps / MPEP 2144 situation.
2. Lim and Konolige independently confirm the same motivation (their abstracts: passive stereo "often fail[s] when the scene does not have enough texture"; "textureless surfaces cannot be matched by stereo, and produce dropouts").
3. Backward-looking — the '981 specification admits the combination: it states "the image processing pipeline application 110 can also illuminate the scene using projected texture and estimate depths of objects within the scene using depth estimation processes similar to those described in U.S. Pat. No. 8,619,082 to Ciurea et al." The applicant thereby concedes that a PHOSITA would pair a projector with a Ciurea-type array pipeline.

Reasonable expectation of success: Both references operate in the same field (triangulation-based depth from image correspondence), use the same underlying disparity/cost-function machinery, and the addition of illumination does not change the geometry of the depth computation. KSR rationales (MPEP 2143) squarely apply:

• (A) Combining known elements (projector + camera array depth estimator) according to known methods to yield the predictable result of denser depth in textureless regions.
• (C) Use of a known technique (Nishihara projected texture) to improve a similar device (Ciurea array) in the same way.
• (F) "Obvious to try": a finite, identified set of projection patterns (random/Nishihara, De Bruijn/Lim, Hamming/Konolige) with predictable results.

Ground 1′ — If Ciurea '082 is disqualified (common ownership, § 102(b)(2)(C))

Substitute a § 102(a)(1)-clean primary reference for the array-camera depth engine — Venkataraman US 2011/0069189 and/or Pain US 2012/0012748 (array-camera architectures; both published before Nov. 18, 2013) — in view of Nishihara + Lim/Konolige, plus the admitted prior art that array cameras perform reference-viewpoint disparity search (the '981 specification itself cites Ciurea '082 as the conventional depth process). The combination logic is unchanged; only the identity of the array-camera reference changes.

---

6. Dependent-claim analysis

Each dependent limitation is met by a further, well-motivated combination. Grouped by theme:

(1) Two 2D arrays in complementary occlusion zones surrounding the illumination system (and the parallel independent claim).

• Combination: Ground 1 + Ciurea '082's own occlusion handling (and/or Ciurea US 9,147,254, "…measuring depth in the presence of occlusions using a subset of images").
• Motivation: Occlusion of the projected pattern by foreground objects is an inherent consequence of co-locating projector and cameras — and Nishihara already places the projector between the two cameras, which literally puts cameras on opposite sides of the projector. Placing multiple cameras on both sides so an occluded region is visible from the other side is the ordinary design response to a known occlusion problem (Ciurea '082 already teaches disregarding occluded pixels and using subsets in which a pixel is visible). Predictable result, no new mechanism.

(2) Each array is a 2×2 (or M×N) array of monochrome cameras.

• Combination: Ground 1 + Venkataraman '189 (array configurations) + routine design.
• Motivation: Monochrome sensors maximize SNR/luminance response for pattern matching; array size is a routine capacity/coverage trade-off. Nothing unpredictable.

(3) Asymmetric projected texture — smaller spatial period in one direction, larger in a perpendicular direction.

• Combination: Lim (patterns defined “over epipolar lines”) + Konolige (blur/phase analysis) + Ground 1; optionally multi-direction structured-light striping (Salvi 2004).
• Motivation: The '981 specification supplies it: where a wide baseline makes one epipolar direction “dominant,” the disparity range along that axis is larger, so the pattern period can be coarser there and finer along the weaker axis. Designing an anisotropic pattern to match anisotropic disparity ranges is a straight optimization of the known Lim/Konolige criteria.

(4) Lower-resolution cameras + at least one higher-resolution camera; higher-res Bayer camera; low-res monochrome; higher-res captures natural texture when projector off.

• Combination: Venkataraman '189 (heterogeneous resolutions) + Nisenzon '760 (π filter group with central Bayer camera) + Ground 1.
• Motivation: Use cheap monochrome low-res cameras for the depth/pattern-matching function and the high-res Bayer camera for color/natural texture — the '981 specification itself cites Venkataraman '189 for exactly this. Turning the projector off to grab natural texture is a trivial time-multiplexing step; Konolige/Lim note active stereo degrades gracefully to passive stereo when the pattern is absent.

(5) Static vs. dynamic illumination; spatial light modulator (LCOS or transmissive LCD).

• Combination: Nishihara/Lim/Konolige (fixed or optimized fixed patterns) → static; digital-projector/Gray-code structured light (Salvi 2004; and the '981's own FIGS. 4E–4I) → programmable patterns via a spatial light modulator. LCOS and LCD SLMs are standard, well-known projector components.
• Motivation: Programmable patterns let the period be tuned to object distance — a recognized precision/defocus trade-off (Konolige's blur analysis supplies the “why”).

(6) Closed-loop adaptive texture — first texture → initial depths → second texture selected from the initial depths; second pattern period finer (at that depth) than the array's spatial resolution there.

• Combination: Ground 1 + Konolige blur/phase analysis + adaptive/gray-code structured light.
• Motivation: The rule that a pattern period must remain resolvable (≈2+ pixels) at the depth of interest, and that finer patterns improve precision but blur out at distance, are known optical constraints (Konolige). Feedback-controlling the pattern to the measured depth is the natural application of those known constraints — “obvious to try” over a finite parameter space. This is the dependent claim with the best (though still modest) non-obviousness argument, because it claims a specific closed-loop control rather than a mere design choice — worth a closer look if litigating.

(7) Array of projectors / projectors projecting different patterns or different spatial periods / controllable.

• Combination: Ground 1 + Gray-code multi-pattern structured light (Salvi 2004) + multiple-projector structured-light systems (e.g., US patents on multi-projector 3D scanning).
• Motivation: Overlaying projectors with successively finer periods is precisely the Gray-code concept (the '981's own FIGS. 4E–4I); using multiple projectors to illuminate different field-of-view regions or to cover the wide projector FOV shown in the '981's FIG. 3E is a routine coverage solution.

(8) Random projected pattern.

• Combination: Nishihara literally discloses a random-dot projected texture. Anticipation-level support for this dependent claim when combined with the array-camera primary reference.

(9) Natural-texture / depth collocation — capture with texture, then without; collocate (common viewpoint, or estimate relative pose from features and reproject, then composite on reliability/confidence).

• Combination: Ground 1 + Ciurea '082 super-resolution and confidence maps + standard SIFT/SURF/BRIEF feature matching and structure-from-motion pose estimation (all conventional by 2013).
• Motivation: Producing a color image registered to the depth map is the ordinary purpose of an RGB-D/point-cloud capture (the '981 specification says as much: “form a point cloud … and/or to generate a mesh and texture”). Compositing depth maps using a confidence map is taught by Ciurea '082 itself; estimating relative pose from common features by minimizing reprojection error is textbook SfM. Predictable.

---

7. Objective indicia (secondary considerations)

• No unexpected results are alleged anywhere in the '981 specification. The specification presents the projector-plus-array system as delivering known benefits (filling textureless regions; more epipolar lines reduce false matches) — i.e., the results are exactly what the artisan would predict from Nishihara/Lim/Konolige applied to a 2D array.
• Long-felt need is weak: the textureless-region problem was identified and addressed in 1984 (Nishihara), and the 2D-array redundancy benefit is a straightforward extrapolation. There is no evidence of a recognized problem that others tried and failed to solve.
• No commercial-success nexus of record tying any alleged success to a specific claimed feature (and much of any such success would flow from unclaimed aspects — the array-camera manufacturing, the π-filter group, etc.).
• Copying/industry praise — none identified.
• Should a patentee assert secondary considerations, the burden shifts to show a nexus to the “different-angle epipolar lines” limitation; the applicant's own admission that this is inherent in a 2D array weakens any such nexus.

---

8. Bottom line and confidence

• Claim 1 is, in my assessment, a strong § 103 candidate — the closest call is not whether the references can be combined (the motivation is express in the references and effectively admitted in the '981 background), but whether the primary array-camera reference (Ciurea '082) is legally available as prior art given the § 102(b)(2)(C) common-ownership concern and its Dec. 31, 2013 issue date.
• Best construction of the case: Ciurea '082 (or, if disqualified, Venkataraman '189 / Pain '148 plus admitted array-camera art) + Nishihara + Lim/Konolige, with additional references for the hardware and adaptive-pattern dependents.
• Strongest dependent claims for the patentee: the closed-loop adaptive-texture claim and possibly the asymmetric-pattern-period claim — these recite specific control/design relationships rather than mere architecture, and merit a more targeted obviousness showing than the claims of record provide.

Confidence:

• High that claim 1 is obvious over the cited references if '082 is available as § 102 art.
• Medium on the § 102(b)(2)(C) availability question for '082 — verify in USPTO Assignment records whether Pelican Imaging (or a common assignee) owned both '082 and the '981 application as of Nov. 18, 2013/2014. This is the linchpin.
• Medium on the exact granted claim wording (claims section truncated in the source I was given).

Recommended verification steps: (1) pull the official granted claims of 10,767,981; (2) confirm '082 ownership as of the '981 effective filing date; (3) run a full-text prior-art search for 2011–2013 "camera array + projector + multi-epipolar depth" to locate a single § 102(a)(1)-clean primary reference; (4) check the family continuations (US 11,486,698; US 12,590,799) for claim-vs-claim differences that would affect a common-scope/OTDP analysis.