US 12298394

Summary of US Patent 12298394B2

Bibliographic data (verified)

Sources: Google Patents (https://patents.google.com/patent/US12298394/en); Justia (https://patents.justia.com/patent/[12298394](/patent/12298394)).

Abstract

The published abstract for the application family reads approximately as follows (the exact granted-patent abstract text was not fully retrievable in my searches, so treat the wording with some caution):

A detector for identifying at least one material property m is described. The detector includes at least one sensor element including a matrix of optical sensors, each having a light-sensitive area, configured to record at least one reflection image of a light beam originating from at least one object, and at least one evaluation device configured to determine the material property by evaluation of at least one beam profile of the reflection image.

The specification describes applying distance-dependent image filters (e.g., depth-from-photon-ratio, depth-from-defocus) and material-dependent image filters (e.g., luminance, spot-shape, speckle/smoothness, grey-level-occurrence contrast/energy/homogeneity/dissimilarity, Law's energy, threshold-area) to a recorded reflection image, to decouple distance information from material classification. Target uses include distinguishing human skin / biological tissue from other surfaces, plus food, metal, plastic, foam, cloth, cosmetics, etc. (Matching text of the family abstract appears at https://portal.unifiedpatents.com/patents/patent/US-20220157044-A1; a continuation-in-family application description appears at https://www.patents-review.com/a/20250334695-detector-identifying-material-property.html.)

Plain-language overview of the independent claims

Important caveat: The full claims set of the granted patent was not fully available in my search results, and the granted claims appear to have been substantially amended during prosecution (the published application US 2024/0241255 A1 was a "Detector"-style application, but the granted patent contains re-cast system and computer-readable medium claims). Based on the Justia claims excerpt, I can reliably describe the following independent claims; there is almost certainly at least one additional earlier-numbered independent claim (e.g., claim 1) whose text I could not retrieve:

• Independent system claim (claim 15, per Justia excerpt): A system with (a) a structured light projector that emits multiple light beams arranged in a first spatial arrangement toward an object; (b) a semiconductor detector (e.g., CMOS, per dependent claim 16) that receives light reflected from the object and outputs an image of the reflected-light intensity; and (c) a processor that: selects at least one reflection region of the image; finds a corresponding reflection region in reference image data derived from multiple reference images; determines the light-intensity distribution of the image's reflection region; compares that distribution with the corresponding reference-region distribution; classifies the object based on that comparison; and outputs a signal carrying the classification information.

  • Plain meaning: A projector-based 3D-sensing/classification system that classifies a material by comparing the shape/intensity distribution of an imaged reflected-light region against learned reference patterns, rather than relying on a single intensity value.

• Independent computer-readable-medium claim (claim 27, per Justia excerpt): A non-transitory computer-readable medium storing instructions that cause a computer system to: receive an image of light reflected from an object illuminated by multiple near-infrared light beams arranged in a first spatial arrangement; select at least one region of the image; determine the intensity distribution of a first subregion that includes the geometric center of the region; determine the intensity distribution of a second subregion extending outward from the first subregion; compare those subregion distributions to the corresponding region in reference image data from multiple reference images; classify the object based on that comparison; and output a signal with the classification.

  • Plain meaning: The software counterpart — center-versus-outer "spot profile" comparison of a reflected structured-light spot against reference data to classify the material.

CAFC 2026 docket search

I searched for CAFC 2026 docket entries tied to TrinamiX GmbH and patent number 12298394 (including an appeal-specific query). No CAFC 2026 appeal or litigation docket referencing this patent was found. Given the patent issued only on May 13, 2025, the absence of a public 2026 Federal Circuit appeal is unsurprising; this does not rule out district-court litigation or an IPR that has not surfaced in the indexed web results, but I have no evidence of any.

Uncertainties

• The granted abstract and complete independent-claim text (notably claim 1) were not fully retrievable; the claim descriptions above cover only the two independent claims visible in the Justia excerpt (claims 15 and 27). For an exact claim 1 recitation, the USPTO Patent Center record for application 18/425,144 would need to be consulted directly.
• The abstract wording I provide tracks the family abstract; minor wording differences from the issued patent's abstract are possible.
• Google Patents' "current assignee" listing shows both Hengen Ingenieurbuero Gbr and TrinamiX GmbH; the recorded assignment chain (as shown in the Google Patents event history) ends with TrinamiX GmbH as assignee, which matches the Justia assignee listing.

I'll search for litigation involving US Patent 12,298,394 across multiple sources.

Initial searches returned no litigation. Let me check litigation-specific databases directly.

Based on my searches of litigation databases and general web sources (Unified Patents portal, CourtListener, UniCourt, Justia, CAFC/PACER-adjacent sources), I found no known litigation involving US Patent 12,298,394.

Search results summary

Conclusion

No litigation is currently known involving US Patent 12,298,394 as of April 26, 2026.

Caveats on confidence:

• The patent issued only on May 13, 2025 (from application 18/425,144, filed January 29, 2024), so the window for filed-and-public litigation is short. Its anticipated expiration is listed as March 13, 2040.
• This search relied on publicly indexed web sources. It does not conclusively rule out: (a) district-court complaints filed but not yet indexed by the searched databases; (b) sealed cases; (c) ITC proceedings; or (d) IPR/PGR petitions, which are often docketed separately and may not surface in general patent-litigation searches. A definitive clearance would require direct PACER queries across all district courts, the ITC EDIS system, and the PTAB's API/PRPS system — none of which returned positive hits in the accessible web sources.

If you'd like, I can run additional targeted searches (e.g., USPTO Patent Center assignment records, PTAB trial searches for IPR2025/IPR2026 petitions against this patent, or district-court-specific queries) to further confirm the absence of proceedings.

Proceedings overview

There are zero (0) AIA trial proceedings — no IPRs, PGRs, or CBMs — on file for US Patent 12,298,394 B2. The USPTO Open Data Portal ("PTAB proceedings on file" block, most recent ingest) returns no trials, and targeted web searches for "12298394 + IPR," TrinamiX IPR/PGR petitions, and PTAB final written decisions referencing this patent likewise surfaced nothing (status breakdown: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution-denied). The defensive bottom line: the patent is entirely untested at the PTAB — every claim remains in force and no estoppel of any kind attaches, so an accused infringer facing assertion today has a clean slate to build the strongest available § 102/§ 103 challenge, but also no prior PTAB record to lean on.

No proceeding-specific sections are required because no proceeding exists; the balance of this memorandum explains that finding, its context, and its strategic meaning.

Strategic summary

No claims are CANCELED, SUSTAINED, or even TESTED. The independent claims (including the re-cast system claim 15 and computer-readable-medium claim 27 visible in the Justia claims excerpt, and any earlier-numbered independent claims such as claim 1 whose full text was not retrievable in my searches) and all dependent claims of US 12,298,394 B2 remain fully in force and have never been the subject of a PTAB petition, institution decision, or Final Written Decision. This is consistent with the patent's age: it issued on 2025-05-13, and even the fastest-filed AIA petitions (theoretical 35 U.S.C. § 315(b) bar notwithstanding, an IPR may be filed before suit) would still be early-stage today. The patent's anticipated expiration is listed as 2040-03-13.

Estoppel landscape — a blank slate. Because no AIA trial has ever been instituted, § 315(e)(2) estoppel does not bar anyone. A defendant currently being asserted against faces no restriction derived from a prior petitioner's record: every § 102/§ 103 ground based on patents or printed publications (and, in a PGR window, § 112 grounds) remains available, subject only to the ordinary § 315(b) one-year-from-service-of-complaint clock and any pre-suit knowledge considerations. Notably, the prior art most relevant to this family is dense — TrinamiX/BASF's own earlier detector patents (e.g., US 10,948,567 B2, US 11,067,692 B2, US 11,698,435 B2, US 11,860,292 B2, and the WO 2014/198,629 and WO 2016/120,392 lineages) plus third-party structured-light/skin-detection art (US 2016/155,006 A1 and US 2016/020,6216 A1 are cited in the specification itself) — so a well-researched petition has ample ammunition.

Pattern signals. There is no repeat-petitioner pattern, no defensive-aggregator involvement (no Unified Patents or similar entity appears in any PTAB or assignment record for this patent), and TrinamiX has never appeared as a PTAB petitioner either. The only enforcement signal is brand-new and outside the PTAB: on 2026-09-03, TrinamiX GmbH (a BASF SE subsidiary) sued [Apple Inc.](/litigations/by-plaintiff/Apple%20Inc.) in the U.S. District Court for the Western District of Texas, Midland Division, alleging that Face ID in iPhone 15/16/17 series and iPad Pro products infringes seven material-and-skin-detection / face-authentication patents (per Reuters and contemporaneous reporting; e.g., https://ca.investing.com/news/stock-market-news/basf-sues-apple-in-us-says-iphones-infringe-face-authentication-patents-[4828257](/patent/4828257)). The reporting does not identify the seven asserted patent numbers, so I cannot confirm whether US 12,298,394 B2 is among them — treat that as unverified. If 12298394 is asserted, Apple's § 315(b) window would run to roughly 2027-09-03, making an Apple IPR (or a joint/aggregator-backed petition) a realistic near-term event to monitor; if it is not asserted, the patent remains a sleeping asset that litigation against Apple may nonetheless make more visible to the PTAB bar.

Recommended next steps

1. Confirm whether 12298394 is in the Apple suit. Pull the complaint in the W.D. Tex., Midland Division action (BASF/trinamiX v. Apple, filed 2026-09-03) from PACER/Public Access to see the seven asserted patents. If this patent is asserted, that filing started the § 315(b) one-year clock for Apple and any real-party-in-interest — calendar the 2027-09-03 IPR deadline immediately.

2. If you are a defendant being asserted against this patent today: the absence of PTAB history is not a weakness in your defense — it is the absence of prior estoppel. Commission a fresh invalidity search now, prioritizing (a) the TrinamiX/BASF detector patent family (same inventors, earlier effective filing dates, some of which are prior art to the 2019 priority date), and (b) the specification's own cited art (US 2016/155,006 A1; US 2016/020,6216 A1; DE 198 46 619 A1; CN 108 363 482 A; US 2018/033,146 A1). Given the patent issued 2025-05-13, a PGR window (9 months from grant, i.e., through roughly 2026-02-13) has already closed, so § 112 and prior-art challenges must now proceed by IPR (§§ 102/103, patents/printed publications only) unless you can show the patent is not entitled to its 2019 priority date — in which case additional art becomes available.

3. Monitor, don't assume. Re-query the USPTO PTAB Open Data Portal API and the PTAB E2E docket search quarterly, and watch the CAFC docket, for any petitions tied to patent number 12298394 or to the seven patents in the Midland, Texas action. If TrinamiX's skin/material-detection claims survive a Markman against Apple and look valuable, an IPR challenge is statistically likely to follow — the absence of any PTAB activity to date is itself a signal that no one has yet been forced to mount one.

Key sources: USPTO ODP "PTAB proceedings on file" block (canonical — zero proceedings); Google Patents bibliographic record (https://patents.google.com/patent/US12298394/en); Justia claims excerpt (https://patents.justia.com/patent/12298394); Reuters via investing.com reporting the 2026-09-03 W.D. Tex. Midland complaint (https://ca.investing.com/news/stock-market-news/basf-sues-apple-in-us-says-iphones-infringe-face-authentication-patents-4828257). No PTAB Final Written Decisions or CAFC appeals exist to cite, and I did not fabricate any.

I'll research the recorded assignment history for US 12,298,394 across USPTO Assignment Center and related sources. Let me run several searches.

Let me search more specifically for the assignment records and reel/frame numbers.

Let me try targeted searches on the USPTO assignment databases directly.

Inventors

All twelve named inventors are listed on the issued patent (US 12,298,394 B2):

Friedrich Schick, Peter Schillen, Patrick Schindler, Andre Schmidt, Michael Eberspach, Christian Lennartz, Robert Send, Lars Diesselberg, Heiko Hengen, Ingmar Bruder, Jakob Unger, Christian Bonsignore.

Employer data at time of filing is only partially determinable from public sources:

• TrinamiX GmbH (Ludwigshafen am Rhein, Germany): The large majority of the inventors have inventor profiles crediting TrinamiX GmbH, including Friedrich Schick and Ingmar Bruder (per PatentLeaderboard inventor-profile pages listing their patent portfolios "at TrinamiX GmbH"). Robert Send, Christian Bonsignore, Jakob Unger, Ingmar Bruder, Michael Eberspach, Christian Lennartz, Andre Schmidt, Friedrich Schick, Peter Schillen and Patrick Schindler are the ten inventors whose interests were recorded as assigned directly to TrinamiX GmbH — consistent with them being TrinamiX employees/inventors under an employment-assignment obligation.
• Hengen Ingenieurbuero Gbr: Inventors Heiko Hengen and Lars Diesselberg recorded their interests first to Hengen Ingenieurbuero Gbr (a German civil-law partnership, "GbR," i.e., an engineering office), which then assigned onward to TrinamiX GmbH. This is consistent with those two inventors having developed the subject matter under an engagement with that engineering firm rather than as direct TrinamiX employees. Note: the recorded assignor name is spelled "DIESSELBERG, Lars" in the USPTO/Google legal-event record while the granted patent names him "Lars Diesselberg"; I have not auto-corrected either spelling.

Unusual pattern: None. There is no evidence of mass inventor departure or a portfolio fire-sale; TrinamiX inventors continue to file and be granted patents through 2025 (per PatentLeaderboard), and the chain terminates at TrinamiX GmbH.

Original assignee

• Entity named on the issued patent: TrinamiX GmbH, Ludwigshafen am Rhein, Germany (original assignee; also the current assignee per the recorded chain and per Justia's assignee listing).
• Line of business: Operating technology company. TrinamiX GmbH develops and commercializes optical sensing technology — near-infrared spectroscopy and mobile material-identification (e.g., scanner solutions integrated into smartphones), 3D depth sensing / beam-profile-analysis systems, and IR sensor products (marketed under the "Hertzstueck™" brand, referenced in the specification itself). The company was spun out of BASF SE and BASF has held a majority stake; this patent family (Detector for identifying at least one material property) is core to TrinamiX's material-identification product line.
• Product embodiment: Yes — the specification and family describe commercial material-identification/3D-sensing technology consistent with TrinamiX's product area (the spec even cites trinamiX's own "Hertzstueck™" IR sensors as usable optical sensors).
• Current status: Operating. No bankruptcy, dissolution, or acquisition of TrinamiX surfaced in any search.

Assignment timeline

Important sourcing caveat: I was unable to retrieve reel/frame numbers and correspondent-of-record fields from the USPTO Assignment Center through the search tools available in this session (targeted queries against assignment.uspto.gov / assignmentcenter.uspto.gov returned no retrievable record pages). The entries below are reconstructed from the Google Patents legal-event feed embedded in the authoritative patent text (three "Assignment of Assignors Interest" events, all recorded 2024-06-20) and the Justia assignee listing. Reel/frame and correspondent fields are therefore marked not retrieved — do not treat as verified. Per the operating rules I have not fabricated them. Direct verification at https://assignmentcenter.uspto.gov/ (search by patent number 12298394) is recommended before relying on the exact reel/frame data.

• Executed date not shown / recorded 2024-06-20 — Reel/frame not retrieved — Google Patents legal event (conveyance type: "ASSIGNMENT OF ASSIGNORS INTEREST")

  • Assignor: Lars Diesselberg (recorded as "DIESSELBERG, Lars"), Heiko Hengen (recorded as "HENGEN, Heiko")
  • Assignee: Hengen Ingenieurbuero Gbr
  • Correspondent: not retrieved
  • Context: The two engineering-firm-affiliated inventors convey their undivided interests to the firm that engaged them — a routine intermediary step in a title-cleaning chain, not an acquisition.

• Executed date not shown / recorded 2024-06-20 — Reel/frame not retrieved — Google Patents legal event (conveyance type: "ASSIGNMENT OF ASSIGNORS INTEREST")

  • Assignor: Hengen Ingenieurbuero Gbr
  • Assignee: TrinamiX GmbH
  • Correspondent: not retrieved
  • Context: Same-day onward assignment consolidating the two inventors' interests into TrinamiX GmbH, the operating company that filed the application.

• Executed date not shown / recorded 2024-06-20 — Reel/frame not retrieved — Google Patents legal event (conveyance type: "ASSIGNMENT OF ASSIGNORS INTEREST")

  • Assignor: Robert Send, Christian Bonsignore, Jakob Unger, Ingmar Bruder, Michael Eberspach, Christian Lennartz, Andre Schmidt, Friedrich Schick, Peter Schillen, Patrick Schindler
  • Assignee: TrinamiX GmbH
  • Correspondent: not retrieved
  • Context: Direct assignment of the ten remaining inventors' interests to TrinamiX GmbH, recorded the same day as the Hengen-chain links — the standard completion of the record before issuance.

Timing context: All three links were recorded 2024-06-20, i.e., after the continuation application 18/425,144 was filed (2024-01-29) and before grant (2025-05-13) — a conventional pre-issuance title-cleaning pattern for an application originally filed in the name of inventors/contractors, ending with the entity that prosecuted the application. No post-issuance assignments are recorded, and no assignment to any LLC, holding company, or third party appears anywhere in the chain.

Timeline diagram

timeline
    title Ownership of US 12298394
    2019 : Priority application filed
    2024 : US continuation 18-425-144 filed
         : Inventors assign to TrinamiX GmbH
         : Hengen firm assigns to TrinamiX GmbH
    2025 : Patent granted to TrinamiX GmbH

NPE / troll-pattern signals

1. Shell-entity transfer — not present. Hengen Ingenieurbuero Gbr is a German civil-law engineering partnership (GbR), not a licensing-only LLC; it transferred its entire interest onward to the operating company on the same day (recorded 2024-06-20). No "IP / Patents / Licensing / Holdings" entity appears in the chain. No registered-agent-address or single-purpose-LLC indicia exist because no such entity is present.
2. Known asserter in the chain — not present. Neither TrinamiX GmbH nor Hengen Ingenieurbuero Gbr matches any public NPE directory (Acacia, Marathon, Intellectual Ventures, Conversant/Wi-LAN, RPX/Unified Patents asserter lists, etc.). No such entity surfaced in any search.
3. Repeat correspondent across the chain — unclear / not determinable. Correspondent-of-record fields could not be retrieved from the Assignment Center in this session, so I cannot assess recurrence. No finding either way.
4. Cascading transfers — not present. Three assignments recorded on one date (2024-06-20) might superficially resemble chaining, but the direction is into the operating company that filed and prosecuted the application (TrinamiX GmbH), not through a sequence of LLCs — a conventional title consolidation, not a cascade.
5. Pre-litigation transfer — not present. No litigation naming US 12,298,394 is known (see prior litigation analysis; no CAFC 2026 docket, no district-court/ITC/IPR records surfaced). The 2024-06-20 recordings occurred during prosecution, roughly eleven months before grant — they cannot be assertion-driven.
6. Bankruptcy fire-sale — not present. TrinamiX GmbH is operating, with no bankruptcy filings; no insolvency-driven patent sale.
7. Privateering — not present. No operating company transferred rights to an NPE for competitor-harassment; no SEC filings or press coverage suggest such an arrangement.
8. Defensive aggregator (anti-NPE) — not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN; it terminates at the original operating assignee, TrinamiX GmbH.

Verdict

Operating-company assertion — with the explicit caveat that no assertion or litigation is currently known. The chain terminates at TrinamiX GmbH, an operating company that develops and ships material-identification/3D-sensing products embodying this technology family; all three recorded assignment links (2024-06-20: two inventors → Hengen Ingenieurbuero Gbr → TrinamiX GmbH, plus ten inventors → TrinamiX GmbH directly) are pre-grant title-cleaning into the entity that prosecuted the application, not transfers to any licensing vehicle. No shell entities, no known asserter, no defensive aggregator, and no litigation appear anywhere in the record, so no NPE pattern is supportable on the available evidence.

Reel/frame numbers and correspondents were not retrievable via the tools used here; verification link for the USPTO Assignment Center search: https://assignmentcenter.uspto.gov/ (search by patent number 12298394).

I'll research the cited prior art on the face of US Patent 12298394 and other relevant references. Let me start with multiple searches.

The Justia excerpt is showing me the cited-reference list. Let me retrieve the full references-cited data from Justia and Google Patents, and identify each citation in detail.

I have to stop taking new tool calls, so I'll synthesize the answer from the verified search results obtained above, plus clearly-flagged reasoning. Below is my prior-art analysis.

Prior-Art Analysis for US Patent 12,298,394 B2 ("Detector for identifying at least one material property")

1. Patent identification — verification against similar numbers

I searched for patent 12298394 and confirm the target is:

• US 12,298,394 B2, title Detector for identifying at least one material property, assignee TrinamiX GmbH, inventors Schick/Schillen/Schindler/Schmidt/Eberspach/Lennartz/Send/Diesselberg/Hengen/Bruder/Unger/Bonsignore.
• Application 18/425,144 (filed 2024-01-29); pre-grant publication US 2024/0241255 A1; granted/issued 2025-05-13; primary examiner Frantz B Jean (per Justia).
• Distinguished from similar family numbers encountered during the search (do not conflate):
  • US 2022/0157044 A1 = family application 17/439,492 (priority 2019-03-14) — a different member of the same family, not the granted patent.
  • US 12,461,238 B2 = a different granted TrinamiX patent in the same family (application 19/041,010) with its own claims.
  • US 2025/0334695 A1 = a later published application in the family (a continuation, not the granted '394 patent).
  • No unrelated "12298394" (e.g., the Orange County debt-collection docket noted in the earlier litigation summary) is a patent.

USPTO-database caveat: The USPTO Patent Center record for 18/425,144 could not be retrieved directly in these web searches; the citations below come from the face-of-patent data mirrored on Justia (patents.justia.com/patent/12298394) and from the specification text on Google Patents. These mirrors reproduce the USPTO "References Cited" list, but for a fully authoritative citation list the USPTO Patent Center/Patent Full-Text database should be checked directly.

2. References cited on the face of US 12,298,394 B2 (as mirrored by Justia)

The Justia references list for the granted patent shows the following U.S. patent documents, foreign patent documents, and WO publications (publication dates as shown on Justia):

Literal-ID caveat (important): The specification of the patent family cites some of these references in truncated form — "US 2016/155006 A1," "US 2018/033146 A1," and "US 2016/0206216 A1." The Justia face-of-patent list shows the corresponding full numbers as US 2016/0155006 A1, US 2018/0033146 A1, and US 2016/0206216 A1. I interpret the in-spec citations as truncations of the full Justia-listed numbers rather than as distinct documents, but I flag this as an interpretation, not a verified identity. The specification also discusses DE 198 46 619 A1 (structured-surface appearance quality determination) as background art; that document did not appear in the truncated Justia reference list I could retrieve, so it may be a background citation rather than an examiner "References Cited" entry.

3. Individual reference analyses

(a) US 2016/0206216 A1 — Kirenko, Device, system and method for skin detection (Koninklijke Philips N.V.)

• Verified: Application 14/995,626, filed 2016-01-14; priority 2015-01-18/19 (EP 15151573.1 and U.S. Prov. 62/104,953); published 2016-07-21; also published as WO 2016/116307 A1 and EP 3247261 A1.
• Content (verified): A device with a thermal sensor input obtaining thermal sensor data of a scene, a light sensor input obtaining light sensor data, and an evaluation unit analyzing both and detecting skin areas (e.g., by temperature range + skin-color range, PPG pulsatility, and/or spatial uniformity of chrominance/intensity). No structured-light projector; no reflection-spot beam-profile/reference-image comparison.
• Relation to the claimed subject matter: Teaches sensor-plus-evaluation-unit classification of skin vs. non-skin from image data — the same classification goal pursued by the '394 patent — but uses thermal + color/PPG cues rather than analysis of a projected structured-light pattern's intensity distribution.
• § 102 anticipation assessment: Weak-to-none as a single-reference anticipator of the independent system claim (Justia claim 15) or the computer-readable-medium claim (Justia claim 27), because it lacks (i) a structured light projector emitting multiple light beams in a first spatial arrangement, (ii) comparison of the light-intensity distribution of a reflection region with reference-image data derived from multiple reference images, and (iii) classification from that distributional comparison. It is a strong § 103 building block for the skin-vs-non-skin classification feature.

(b) US 2016/0155006 A1 — Makkapati et al., Device and method for skin detection (published 2016-06-02 per Justia)

• Content (as described in the '394 specification): A device with an illumination unit configured to project a predetermined illumination pattern onto a scene, an imaging unit acquiring an image of the scene, and an evaluation unit that evaluates the acquired image by analyzing the imaged illumination pattern as reproduced in the image to detect skin areas and distinguish them from non-skin areas.
• Relation to the claimed subject matter: This is structurally the closest of the U.S. citations to the granted independent claims: it actively projects a pattern, images the scene, and classifies (skin/non-skin) by analyzing the imaged pattern. The '394 claims add (per the Justia excerpt) that the projector emits multiple light beams in a first spatial arrangement, that a reflection region is matched against reference image data derived from multiple reference images, and that intensity-distribution comparison drives the classification.
• § 102 anticipation assessment: Potentially the strongest single-reference candidate, but full anticipation of claims 15 and 27 is doubtful on the retrieved claim text because Makkapati's pattern analysis is not described as comparing the center/outer intensity distribution of an imaged spot against reference image data accumulated from multiple reference images, and its classification target is limited to skin. I cannot rule out anticipation of an earlier, broader independent claim (the full text of claim 1 was not retrievable) without the complete claim set; if claim 1 is drawn broadly to "project a pattern, image the reflected pattern, classify the object from the imaged pattern," Makkapati would be a serious § 102 candidate.

(c) US 2018/0033146 A1 — Bleyer et al., Systems and methods for determining a depth map and a reflectivity map from a structured light image (Microsoft; published 2018-02-01)

• Content (verified from the '394 specification and public record): Captures a structured-light image (dot pattern), determines a depth map by triangulation based on the dots, and determines a reflectivity map based on the depth map and additional analysis of the dots in the captured structured light image.
• Relation to the claimed subject matter: This is the closest art to the core technical operation of the granted claims — per-dot analysis of a structured-light image beyond mere position (reflectivity/intensity), tied to depth. However, Bleyer derives a reflectivity map for scene geometry; it does not classify a material property by comparing a spot's intensity distribution with stored reference images of known materials.
• § 102 anticipation assessment: Moderate as to the sensor/projector/processor framework and the "analyze dot intensity" step; weak as to the final classification-by-comparison-with-reference-data step. If the granted independent claims were amended to require the explicit "compare with reference image data from multiple reference images → classify → output classification signal" sequence (as the Justia excerpt of claims 15 and 27 indicates), Bleyer does not anticipate those claims as a single reference. It is the most dangerous § 103 combination anchor, however, since it supplies the structured-light + intensity-analysis hardware.

(d) US 2017/0363465 A1 — Send et al. (TrinamiX; published 2017-12-21)

• Content: A TrinamiX optical detector application from the same inventive community, disclosing the depth-from-photon-ratio (DPR) and related beam-profile-analysis techniques for determining longitudinal (distance) coordinates of an object from a light spot on a sensor matrix.
• Relation to the claimed subject matter: The '394 specification's distance-feature machinery (φ1z via depth-from-photon-ratio / depth-from-defocus filters) builds directly on this line of TrinamiX work.
• § 102 anticipation assessment: Weak for the granted classification claims, because distance measurement ≠ material classification via reference-image comparison. It is highly relevant as § 103 prior art for the distance-decoupling features that appear in dependent claims of the family.

(e) US 2018/0276843 A1 — Send (TrinamiX; published 2018-09-27)

• Content: Another TrinamiX detector application (position/distance determination via optical sensing) in the same DPR/beam-profile family as (d).
• § 102 anticipation assessment: Weak for the same reason as (d). Flag: I could not independently verify the exact title of US 2018/0276843 A1 in the available snippets; treat the title as unverified.

(f) CN 108363482 A (published August 2018)

• Content (per the '394 specification): Method for controlling smart televisions via three-dimensional gestures using binocular structured light — synchronously acquiring images with a binocular camera, reconstructing a 3D image, preprocessing, segmenting/recognizing 3D gestures, and converting them into operation instructions.
• § 102 anticipation assessment: None-to-weak for material-property claims: it classifies gestures, not materials, and uses binocular triangulation rather than single-camera intensity-distribution comparison against reference image data.

(g) WO 2014/097181 A1 (published 2014-06) and (h) WO 2014/198629 A1 (published 2014-12) — TrinamiX

• Content: The '394 specification expressly cites these in connection with (i) transversal sensors and position-sensitive devices (WO 2014/097181 A1) and (ii) deriving transversal position from the light spot position on a matrix of optical sensors (WO 2014/198629 A1). Both are foundational TrinamiX optical-position-detector applications.
• § 102 anticipation assessment: Weak for material-classification claims; relevant to sensor-matrix/position-detection limitations and to § 103 combinations with the DPR/DFD features.

(i) WO 2012/110924 A1 (published 2012-08)

• § 102 anticipation assessment: Cannot be characterized with confidence. I could not verify the title, applicant, or content of this document in the available search results. It predates the family's earliest work and may be an early TrinamiX/BASF optical-detector application, but I will not guess its content. No anticipation mapping is offered pending verification of the document.

(j)–(l) WO 2018/091638 A1, WO 2018/091649 A1, WO 2018/091640 A1 (TrinamiX; 2018 publications)

• Content: Almost certainly TrinamiX PCT applications on optical detection/beam-profile analysis (the WO numbering 2018/0916xx indicates PCT publications in the week of ~2018-05-24). Flag: Justia lists WO 2018/091640 A1 as August 2018, which is inconsistent with the 2018/0916xx publication-number convention; I could not resolve this discrepancy and list the date as shown. Titles/content of each individual WO document were not verified in my searches.
• § 102 anticipation assessment: Weak as single-reference anticipators of the classification claims; relevant as § 103 art on distance-dependent beam-profile analysis from the same family.

(m) WO 2019/042956 A1 (published 2019-03-07)

• Content: A TrinamiX PCT application published just eight days before the '394 family's earliest priority date (2019-03-15; sibling records show 2019-03-14). Flag: exact title/content not verified in my searches.
• § 102 anticipation assessment: It is technically prior art under AIA § 102(a)(1) (published before the priority date), but without verified content I cannot map it to specific claims. Given the publication timing and assignee, it is likely a same-community optical-detector application and thus primarily § 103-relevant.

(n) DE 198 46 619 A1 (applicant-cited background; publication 1999 class)

• Content (per the '394 specification): Equipment for determining structured-surface appearance quality — evaluating electrical measurement signals from a photosensor array to derive a structure code characterizing the measurement surface.
• § 102 anticipation assessment: None for the structured-light/classification claims; it is surface-texture metrology background art.

4. Summary ranking of § 102 anticipation potential

Given the limitations noted (full claim 1 text and full dependent-claim set not retrievable; only independent claims 15 and 27 recoverable from the Justia excerpt), my claim-level conclusions are necessarily provisional. Against the retrieved independent claims (system claim ~15 and computer-readable-medium claim ~27), the common inventive requirements are: (1) structured-light projector emitting multiple light beams in a first spatial arrangement; (2) semiconductor image sensor (CMOS) receiving reflected light; (3) selecting a reflection region; (4) finding a corresponding reflection region in reference image data derived from multiple reference images; (5) determining and comparing the light-intensity distribution of the region (center subregion vs. outward subregion) with the corresponding reference-region distribution; (6) classifying the object based on the comparison; (7) outputting a signal carrying the classification.

Bottom line: No single cited reference appears, on the retrieved record, to disclose the complete combination of the granted independent claims — most critically the classification of the object by comparing the intensity distribution of the imaged reflection region against reference image data derived from multiple reference images, followed by output of a classification signal. The closest anticipatory threats are Makkapati (US 2016/0155006 A1) on the classification concept and Bleyer (US 2018/0033146 A1) on the structured-light intensity-analysis architecture. A definitive § 102/§ 103 opinion would require (a) the complete granted claim set from USPTO Patent Center for application 18/425,144 (notably the full text of claim 1 and the dependent claims), and (b) verification of the unverified WO documents (WO 2012/110924 A1 and WO 2019/042956 A1 in particular).

Obviousness Analysis under 35 U.S.C. § 103

U.S. Patent 12,298,394 B2 — "Detector for identifying at least one material property"

Important scope note up front. The prior-art subsection of the Google Patents page (the examiner-cited "Patent Citations" / "Non-Patent Citations" lists) was not fully retrievable in the materials available to me, and the complete text of claim 1 was not captured (claims 2–14, 16–26 clearly depend from system claims 1/4/15 and a near-IR/randomized/DOE/VCSEL feature set, per the Justia excerpt at https://patents.justia.com/patent/12298394). My element-level analysis below therefore uses the two independent claims whose text I can quote with confidence — claim 15 (system) and claim 27 (computer-readable medium) — and the prior art expressly discussed in the specification's background section, supplemented by the closely analogous structured-light classification reference US 9,436,868 B2 surfaced in my searches. Where a conclusion depends on claim text I could not verify, I say so explicitly.

1. Claim scope used for the analysis

Claim 15 (system):

1. A structured light projector emitting a plurality of light beams arranged in a first spatial arrangement relative to each other toward an object;
2. a semiconductor detector receiving light reflected from the object and outputting an image characterizing an intensity of the reflected light; and
3. a processor configured to:
   • select at least one reflection region of the image;
   • determine a corresponding reflection region in reference image data derived from a plurality of reference images;
   • determine a distribution of light intensity of the reflection region of the image;
   • compare that distribution to the distribution of light intensity of the corresponding reference region;
   • determine a classification of the object based on that comparison;
   • output a signal carrying the classification information.

Claim 27 (non-transitory computer-readable medium) tracks the same concept in method/software form, adding: light beams emitted in a near-infrared range, and comparison of the intensity distribution of a first subregion inclusive of the geometric center of the region and of a second subregion extending from the first subregion against the corresponding reference region.

The specification (authored description, "the invention" section) explains the technical problem the claims address: "most image filters used for beam profile analysis (BPA) yield features that depend on both, distance and material such that reliable identification and classification of the material is not possible. A typical example is the width of a beam profile reflected from a translucent material." The claimed solution compares reflected structured-light spot intensity distributions (center vs. outer subregions) against reference distributions derived from multiple reference images, and classifies accordingly — decoupling material information from distance information.

2. Person of ordinary skill in the art (PHOSITA)

A PHOSITA at the March 2019 priority date would be an engineer or scientist with a B.S./M.S. (or equivalent experience) in electrical engineering, optical engineering, applied physics, or computer science, and roughly 2–5 years of experience developing structured-light 3D sensors, CMOS/CCD imaging systems, near-infrared illumination, and image-based pattern/object classification. The person would be conversant with: DOE-based dot projectors, camera calibration and triangulation, depth-from-defocus and photon-ratio range techniques, image filtering (blob detection, edge/contrast statistics, texture/co-occurrence measures), and classifier training against reference/database imagery. The patent itself presupposes such a person by defining "passes a hypothesis testing" for material filters and by referencing commercial infrared sensors and known DPR/DFD algorithms.

3. Prior-art universe (as acknowledged in the patent and confirmed by search)

The specification's background discussion identifies the following closest prior art (the "Prior Art" section of the page):

Collectively this art already contains, by March 2019: (i) structured-light dot projectors + CMOS imaging (US 2018/033146; CN 108 363 482); (ii) material classification from an imaged illumination pattern in reflective mode (US 2016/155006 skin/non-skin); (iii) object classification from structured-light intensity-derived measurements compared against stored class references (US 9,436,868); (iv) spatial "structure code" extraction from photosensor-array intensity distributions (DE 198 46 619); and (v) center-vs.-edge beam-profile signal splitting (the DPR/DFD art the applicant itself characterizes as known).

4. Element-by-element mapping (claim 15)

5. Lead combinations and the § 103 rationale

Under Graham v. John Deere and KSR Int'l v. Teleflex, the question is whether the differences are such that the claimed subject matter as a whole would have been obvious to the PHOSITA, with the combination motivated by a reasonable expectation of success — not by a rigid TSM test.

Combination 1 (strongest for claim 15): US 2018/033146 A1 + US 2016/155006 A1 (+/- US 9,436,868 B2)

• What each contributes. US 2018/033146 supplies the full claimed front end: a structured-light projector casting a dot field, a camera forming an image of the reflected dots, and per-dot intensity analysis (its reflectivity map). US 2016/155006 supplies the material-classification goal and mechanism in the identical optical geometry: it teaches that the way a projected illumination pattern is imaged after reflection carries skin-vs.-non-skin material information, and provides an evaluation unit to extract that information. US 9,436,868 supplies the classification backend: object type is determined by comparing structured-light intensity-derived measurements against stored per-class reference ranges and outputting the class.
• Why combine. The references are in the same field (optical material identification and structured-light sensing) and solve complementary parts of the same problem — US 2018/033146 even computes reflectivity (a material property) but stops short of labeling; the patent's own background complains that reflective-mode material identification is normally done at wavelengths above 1000 nm or in transmission, which is precisely the problem US 2016/155006 (skin detection in reflective projected-pattern mode) and US 9,436,868 (translucence-based classification) had already attacked. A PHOSITA seeking "reliable identification of materials from a reflection beam profile" would have had an express design incentive to bolt the pattern-analysis/classification teachings of US 2016/155006 and/or US 9,436,868 onto the structured-light reflectivity pipeline of US 2018/033146.
• Predictability / obvious-to-try. The number of candidate material-descriptive metrics known in the art (reflectivity, translucence, speckle/smoothness, structure code) is small and enumerated in the prior art; substituting "compare the region's intensity distribution against reference distributions" for "compare scalar reflectivity/translucence against threshold ranges" is a design choice among a finite set of known alternatives, yielding a predictable result. Using a plurality of reference images to form the comparison basis is the textbook way to build any classifier/database and is disclosed generically in the patent's own "database of pre-determined material properties from test measurements" embodiment.

Combination 2 (strongest for claim 27): DPR/DFD center-vs-edge art (WO 2014/097181 A1 family; the spec's own admitted "depth-from-photon-ratio filter") + material classification (US 2016/155006 / US 9,436,868)

• Claim 27's center-subregion/second-subregion intensity-distribution split is, structurally, the DPR "center information vs. edge information" beam-profile division the specification itself describes as a known distance filter (integrate/sum inner vs. outer areas, divide, compare). In the prior art this split was used for distance; the only arguably new twist is using the same split for material classification by comparing to reference distributions.
• Why combine / why obvious. Once US 2016/155006 teaches that imaged-pattern analysis classifies skin vs. non-skin, and US 9,436,868 teaches comparing structured-light intensity measurements to class references, applying a known center-vs.-outer region decomposition to obtain a material-sensitive feature vector is a straightforward adaptation with a reasonable expectation of success — the PHOSITA knows from the DPR art that the center/edge ratio of a back-scattered spot encodes sub-surface scattering behavior, and the skin-detection art already ties subsurface scattering/translucency to material class. One would arrive at "compare center and outer intensity distributions of the reflected spot to reference data; classify; output" without invention.
• DE 198 46 619 A1 adds further support: photosensor-array signals were already being converted into surface "structure codes," evidencing that array-level intensity-distribution statistics were a routine classification input.

Combination 3 (supplementary): US 2018/033146 A1 or CN 108 363 482 A + DE 198 46 619 A1

A structured-light depth/reflectivity system (US 2018/033146) or a structured-light 3D + classification pipeline (CN 108 363 482) combined with the array-signal "structure code" evaluation of DE 198 46 619 would render the select region → derive distribution → classify core of claim 15 obvious: DE 198 46 619 demonstrates that an imaged intensity distribution over a sensor array can be coded into a surface/material-classifying descriptor, and CN 108 363 482 demonstrates the full "structured light → image → segment → classify → output command" loop.

6. Where the obviousness case is weakest (likely basis for allowance)

The examiner (Frantz B Jean) allowed the claims after the applicant re-cast them into the distribution-comparison system/medium form, so the record presumably turned on these gaps:

1. "Distribution" vs. scalar metrics. US 9,436,868 compares scalar values (background intensity, fitted maximum, reflectance); US 2018/033146 derives scalar per-dot reflectivity. Neither teaches comparing the full spatial intensity distribution of a reflected region — let alone center vs. outer subregion distributions — against reference-image distributions to classify material. If "distribution" is given its full weight (shape, gradients, center-to-edge roll-off), the combination analysis above requires impermissible hindsight to supply the missing limitation.
2. No teaching that spot-profile shape separates material from distance. The specification's central problem statement — BPA features (e.g., spot width for translucent materials) are confounded by distance — cuts in the patentee's favor: the cited art does not recognize, let alone solve, the distance/material confounding, and none of US 2016/155006, US 2018/033146, or US 9,436,868 discloses distance-normalizing or distance-decoupling the intensity distribution before material classification. A challenger must show a PHOSITA would have had a reasonable expectation that distribution-shape comparison alone would classify reliably across varying distances — a showing not present in the identified art.
3. Reference-image-data limitation. The "reference image data derived from a plurality of reference images" language is narrower than "threshold ranges" (US 9,436,868) or a pattern-geometry skin rule (US 2016/155006); unless the challenger can point to an explicit disclosure of multi-image reference matching, this element may survive as the claim's keystone.
4. Secondary considerations are unknown in the available record — no litigation (see prior analysis) and no objective indicia located — so the analysis rests on the technical merits alone.

7. Bottom line

On the available record, the most plausible § 103 challenge would allege that claim 15 is obvious over US 2018/033146 A1 in view of US 2016/155006 A1 and/or US 9,436,868 B2, and that claim 27 is obvious over the same art further in view of the admitted DPR/DFD center-vs-edge beam-profile art (WO 2014/097181 A1 family), with DOE dot-projector and multi-image classifier training treated as common general knowledge. The motivation story is strong for the apparatus and pipeline elements (same field, complementary teachings, known design incentives, finite set of known material-descriptive features). The outcome would hinge on two disputed limitations for which I found no explicit prior-art disclosure: (i) comparing intensity distributions (center vs. outer subregions) rather than scalar intensity-derived values, and (ii) classification against reference image data built from a plurality of reference images. Without a reference teaching those, a court would likely credit the applicant's distance/material-decorrelation problem as the non-obvious contribution; with such a reference (e.g., an undiscovered structured-light texture- or speckle-classification patent from the 2010s not yet indexed in my searches), the claims would be vulnerable.

Evidence limitations to flag: (a) full claim 1 text and claims 2–14 details were not retrievable, so claim 1 may contain additional narrowing features strengthening validity; (b) the USPTO examiner-cited references and any Reasons for Allowance were not accessible, so I cannot rule out that the cited art already includes a closer distribution-comparison reference than those analyzed here; (c) no PTAB, district-court, or CAFC record exists as of April 26, 2026 to test these theories adversarially (per the litigation search above).