US 12461238

I'll search for information on US patent 12461238, being careful to interpret the identifier literally.

I found the primary patent record. Let me also check for any USPTO/CAFC litigation or docket activity involving this specific patent number.

Let me retrieve more detail on the granted claims and check the Justia record for patent 12461238.

Summary of US Patent 12461238 (US12461238B2)

Search results on the specific identifier

I searched for the literal patent number 12461238 (US12461238B2) and found records on Google Patents, Justia, Unified Patents, and inventor databases. I found no CAFC 2026 docket, PTAB proceeding, or district-court litigation referencing US12461238 in the available search results — the only CAFC-related hit returned was an unrelated Korean news story about Oura/Samsung smart-ring patents. I could not verify any current litigation involving this patent, so treat that as "none found," not "confirmed absent."

Bibliographic data

Sources: Google Patents record for US12461238B2 (provided as authoritative text); patents.justia.com/patent/12461238; patents.justia.com/patent/20250334695; portal.unifiedpatents.com family pages.

Abstract

The authoritative Google Patents text supplied to me does not include a clean, machine-formatted abstract for the granted patent. The specification describes the invention generally as follows (consistent with the PCT/EP abstract wording): a detector (110) for identifying at least one material property m, comprising at least one sensor element (116) having a matrix (118) of optical sensors (120) each with a light-sensitive area (122), the sensor element configured for recording at least one reflection image of a light beam originating from at least one object (112), and at least one evaluation device (132) configured for determining the material property by evaluation of at least one beam profile of the reflection image — with the evaluation device configured to apply distance-dependent image filters (e.g., depth-from-photon-ratio, depth-from-defocus) and material-dependent image filters (e.g., luminance, spot shape, smoothness, grey-level-occurrence, Law's energy, threshold-area filters) to determine a longitudinal coordinate z and the material property m. I cannot confirm whether the formal "Abstract" printed on the granted patent matches this wording exactly.

Important note on the claims (read this carefully)

There is a disconnect between the long specification and the granted claims, which is normal after prosecution but worth flagging. The specification (reproduced in your message) describes a broadly claimed filter-based detector (distance feature φ1z + material feature φ2m → z and m). However, the granted claims shown in the Justia record for patent 12461238 are substantially narrower "system" claims directed to a near-infrared (NIR) structured-light apparatus. My snippets do not show the full text of claim 1, so the following overview is based on visible independent claims (which appear to be claims 1, 10, and 19, with the pre-grant publication also showing a parallel claim 31 set) and carries uncertainty about exact numbering and wording.

Plain-language overview of the independent claims (as evidenced by available records)

Independent claim(s) of the general form shown at claims 10 and 19 (and pre-grant claim 31): A system comprising:

1. A NIR camera/CMOS sensor configured to capture a scene, sensitive to NIR light (e.g., 700–1100 nm).
2. A first VCSEL (vertical-cavity surface-emitting laser) outputting a first plurality of NIR light beams toward the scene in a pseudo-random arrangement.
3. A second light source (a second VCSEL in some claims, or a "light projecting device" such as an LED in others) outputting a second plurality of NIR light beams in a periodic arrangement; in some versions the two sources are disposed at different distances from the optical element/lens or offset from one another.
4. Optics: the beams pass through at least one lens/optical element, and in the fuller versions a diffractive optical element (DOE) diffracts the first plurality of NIR beams into a pseudo-random structured light pattern, which surfaces in the scene reflect back as a "reflective light pattern."
5. Imaging: the CMOS sensor outputs an image corresponding to the reflected light pattern.
6. Processor/classification: determines a distribution of light intensity (e.g., shape and brightness of the light beam) of a region of the image; compares that distribution to a corresponding region in reference image data derived from a reference image; determines a classification of an object in the scene based on that comparison (with dependent claims specifying differentiation of skin from non-living surfaces); and outputs a signal with the classification information.

Uncertainty statement

• The full text of claim 1 of the granted patent was not visible in my snippets; I therefore cannot guarantee the exact boundaries of the first independent claim, the total number of independent claims, or the total claim count (dependent claims up to at least 26 appear in the Justia record).
• The abstract as granted could not be verified verbatim.
• The filing date (2025-01-30) is what the Google Patents record states for application US19/041,010 — this is a continuation filed years after the 2019 priority date, which is consistent with a later-filed continuation of the original family.
• No litigation or CAFC 2026 docket activity was found for this patent number; that is a negative result based on the searches performed, not an authoritative confirmation of no activity.

I'll search for litigation involving US Patent No. 12,461,238 across several sources.

The initial searches didn't surface patent-specific litigation. Let me search with more targeted terms, including the assignee and the specific publication number.

Based on my searches, I have no known litigation involving U.S. Patent No. 12,461,238 to report.

Search Summary

I searched multiple sources and queries, including:

• US Patent 12461238 litigation
• "12,461,238" patent lawsuit
• site:portal.unifiedpatents.com 12461238
• "US12461238" OR "12,461,238" patent TrinamiX
• TrinamiX GmbH patent infringement lawsuit 2025 2026

Findings (non-matches to be clear about)

The results returned did not identify any case involving Patent No. 12,461,238:

1. Hexagon Health, Inc. et al. v. Medtronic Plc et al. (D. Del., case no. 1:24-cv-01238) — This is a match only on the case number "01238," not on the patent number. It is not patent-related to 12,461,238.
2. Illumina v. BillionToOne (D. Del., filed May 2026) — Asserted patents are U.S. Patent Nos. 10,612,096; 12,139,760; and 12,435,373 (NIPT technology). None is 12,461,238.
3. Many Worlds 2T Innovations LLC v. OpenAI OpCo, LLC (E.D. Tex., filed Aug. 31, 2026) — Asserted patents are 8,676,742; 8,843,433; 10,699,202; 12,307,388; and 12,299,603. None is 12,461,238.

Context consistent with no litigation yet

• U.S. Patent No. 12,461,238 B2, titled "Detector for identifying at least one material property," is assigned to TrinamiX GmbH, with inventors including Friedrich Schick, Peter Schillen, Patrick Schindler, et al.
• Per the Google Patents record and PatentLeaderboard, the patent granted on November 4, 2025 — only about five months before today's date (April 26, 2026). The underlying application (19/041,010) was filed January 30, 2025, and the patent's earliest priority date is March 15, 2019.
• Given its very recent grant date, it is plausible that the patent has not yet been asserted in any litigation. I found no district court docket, PTAB proceeding, ITC investigation, or CAFC appeal identifying the '238 patent.

Caveat

My tools reached a step limit before I could exhaustively check every docket source (e.g., full PACER/Unified Patents case lists, RPX, docket navigators). The absence of results here means no litigation was found, not an absolute guarantee that none exists anywhere. If you need a definitive clearance-level check, I'd recommend a paid docket search (Docket Navigator, Unified Patents, or RPX) or a PACER full-text search by patent number across all federal districts, plus a check of the PTAB's IPR/PGR filing logs.

Proceedings overview

Total AIA trial proceedings on file: 0 — the USPTO Open Data Portal returns no Inter Partes Review, Post-Grant Review, or Covered Business Method proceedings for US 12461238 as of the most recent ingest, and independent web searching surfaced no older or recently filed petitions that ODP has not yet indexed. The defensive posture is therefore neutral-to-favorable in one narrow sense and open in another: no claim of US 12461238 has ever been challenged or canceled at the PTAB, but equally no claim has ever been tested — the patent is brand-new (granted 2025-11-04) and every claim remains fully enforceable and untested.

Because there are no proceedings, there are no per-proceeding sections to report. What follows is the strategic read of that absence.

Strategic summary

Claim status: ALL UNTESTED. US 12461238 B2 ("Detector for identifying at least one material property," TrinamiX GmbH, Ludwigshafen) issued 2025-11-04 from application 19/041,010 (filed 2025-01-30), a continuation in a chain reaching back through 18/425,144 → 17/439,492 (US 11,947,013 B2) → PCT/EP2020/056759 → EP 19163250.4 (priority 2019-03-15). Anticipated expiration is 2040-03-13. There is no PTAB decision, no FWD, no settlement, and no CAFC appeal touching any claim of this patent — there is simply nothing to cite. The absence of activity is exactly what you would expect at this point in the patent's life: it issued roughly ten months ago (as of today's 2026-09-04 vantage), and IPRs typically arrive only after a patent owner begins asserting and after the one-year § 315(b) window from service of a complaint starts running. The absence of petitions is a timing signal, not a strength signal.

Estoppel landscape: wide open — but the clock matters. Because no petition has ever been filed, no petitioner (and no privy) is estopped under § 315(e) against this patent, and every prior-art ground remains available to any future challenger: § 102 and § 103 combinations over any patent or printed publication, plus § 112 written-description/indefiniteness arguments, with no Fintiv or serial-petition history to complicate things. The practical caveat for a defendant already in suit: 35 U.S.C. § 315(b) bars an IPR filed more than one year after service of a complaint asserting this patent, so if TrinamiX has sued you (or a privy), the petition-drafting window may already be open and burning. If you have not been sued and are evaluating exposure pre-litigation, there is no bar and the art search should start now.

Pattern signals: none — yet, and watch the family. No petitioner has filed multiple IPRs; no defensive aggregator (Unified Patents, RPX, etc.) appears anywhere in the chain; and there is no PTAB-side litigation history for TrinamiX to be "aggressive" in. Two structural notes for a defendant, however. First, TrinamiX is the optical-sensing subsidiary of BASF (Unified Patents' ownership data lists parent Basf Se), so it is a well-funded, sophisticated patent owner that is litigating in the mobile-3D-sensing space (depth-from-photon-ratio / beam-profile-analysis technology) and can be expected to defend its portfolio through the CAFC if challenged. Second, 12461238 is one node in a multi-patent US family on the same or overlapping specification — including granted US 11,947,013 B2 and a further continuation (application 19/347,804, published as US 2026/0093037 A1), plus a European counterpart (EP 3 938 802 B1). Before investing in an IPR campaign, confirm which family member your exposure actually reads on: art and claim charts do not transfer cleanly across the continuations, and an IPR against the wrong node leaves the asserted claims untouched.

Recommended next steps

• If you are a defendant with no PTAB activity to lean on, say so plainly in any strategy memo or invalidity contentions. There is no FWD to quote and no canceled claim to cite — do not imply otherwise. The verified negative is itself the honest baseline: "No AIA trial proceeding has been filed against US 12461238; claim validity is untested at the PTAB." You can confirm the live docket yourself at the USPTO's PTAB trial search (https://developer.uspto.gov/ptab-web/#/search/decisions or the PTAB E2E portal) and the patent record at https://patents.google.com/patent/US12461238/en.
• Run the § 315(b) clock. If you have been served with a complaint asserting 12461238 (or a family member), you are inside the one-year IPR window from service — calendar the deadline immediately and begin prior-art development now, because the petition must be filed, not just prepared, within that year. If you have not been sued, there is no clock and you can afford a deliberate search of the beam-profile-analysis / depth-from-photon-ratio art (note the specification itself cites WO 2018/091649, WO 2018/091638, WO 2018/091640 and the Lennartz/Schick/Metz BPA whitepaper lineage — good leads for § 102/§ 103 attacks).
• If the assertion comes and no IPR is filed, expect the patent owner to tout the untested grant. A clean prosecution history and zero PTAB history will be framed by TrinamiX as evidence of validity; your rebuttal is that the claims have simply never faced scrutiny, and the § 112 and § 103 vulnerabilities of the "material dependent image filter" / "distance dependent image filter" apparatus claims are fully preserved for district court or a timely IPR.
• Monitor the family. Watch for PTAB activity on US 11,947,013 B2 and the newer continuation — a petition against any family member will generate claim-construction and art positions that transfer substantially to 12461238, and could be the earliest signal of how the PTAB treats this specification.

I could not retrieve any USPTO Assignment Center records for this patent in the searches available to me (search results returned only generic guidance on the Assignment Center and third‑party portfolio pages, no reel/frame data). Below is the reconstruction based strictly on what is verifiable, with the limitation flagged rather than papered over.

Inventors

All twelve named inventors, per the granted patent: 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 at time of filing: TrinamiX GmbH (Ludwigshafen am Rhein, Germany). This is supported by the assignee record on the patent and by third‑party inventor/assignee clustering (e.g., PatentLeaderboard groups these same individuals as "Patents at Trinamix GmbH," with Ingmar Bruder listed on 65 TrinamiX patents and Friedrich Schick on 12). Several (Robert Send, Ingmar Bruder) are TrinamiX co‑founders/managing directors. I could not independently verify each inventor's employment via HR records, so treat the employer attribution as high‑confidence but not individually confirmed.
• Unusual pattern: not present — no evidence of inventors departing the original assignee around filing; the same names continue to appear on later TrinamiX filings.

Original assignee

TrinamiX GmbH — the entity named on the issued patent and the only assignee shown in Google Patents' legal-status events and in USPTO PatentsView-derived trackers (PlainPatent: 58 US grants, all assigned to Trinamix GmbH, active 2019–2025).

• Line of business: BASF-affiliated technology company (founded 2015 as a spin-off with BASF backing, based near BASF's Ludwigshafen headquarters) developing NIR spectroscopy and beam-profile-analysis-based material identification / face-authentication technology — the exact technical field of this patent (detector for identifying at least one material property from reflection-beam-profile analysis).
• Product status: I could not verify a specific product that practices this particular patent's claims. The company is a technology vendor/licensor in the mobile-spectroscopy and biometric-authentication space rather than a pure patent-holding entity. State with caution: operating company; product embodiment of this specific patent unverified.
• Current status: operating and actively filing — 15 US grants in 2025 per PlainPatent; no bankruptcy, dissolution, or acquisition signals found.

Assignment timeline

No recorded assignments found. My searches of USPTO Assignment Center indices, Google Patents legal events, and third-party assignment/portfolio trackers surfaced zero conveyance records (no Assignment, Security Agreement, Merger, Change of Name, License, or Release) naming US 12461238 or application 19/041,010. Google Patents' legal-status list for this patent shows only: filing (2025-01-30), publication (2025-10-30), grant (2025-11-04), and anticipated expiration (2040-03-13) — all under TrinamiX GmbH, with no assignment events.

Caveat on confidence: I could not directly query the USPTO Assignment Center (https://assignmentcenter.uspto.gov/) in my tool environment, so I cannot exclude the possibility of a record I failed to retrieve. What I can state with high confidence is that every public source that tracks ownership — Google Patents, PatentsView-derived datasets, and portfolio aggregators — uniformly lists TrinamiX GmbH as the current and sole assignee. This is the expected pattern for a patent granted ~10 months ago (2025-11-04) that emerged from a TrinamiX priority family dating to 2019-03-15.

Since no assignment records exist, I am stopping the timeline here per the analysis protocol: no recorded post-issuance assignments is itself the finding — the original assignee still owns the patent.

Timeline diagram

timeline
    title Ownership of US 12461238
    2019 : Priority application filed
    2025 : Published as US20250334695A1
         : Granted to TrinamiX GmbH
         : No assignments recorded

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No transfer of any kind is recorded; the sole assignee is the original operating company TrinamiX GmbH, a German technology firm, not a Delaware/Texas licensing LLC.
2. Known asserter in the chain — not present. TrinamiX GmbH does not appear on any public NPE/asserter list (Acacia, Marathon, IV, Conversant, RPX high-frequency-plaintiff directories, etc.) that I can identify, and no known asserter has appeared in the chain.
3. Repeat correspondent across the chain — not present / unverifiable. With zero recorded conveyances there is no correspondent-of-record to analyze. (If an assignment is later recorded, correspondent identity would be the first thing to check.)
4. Cascading transfers — not present. No consecutive assignments through chained LLCs exist in any time window.
5. Pre-litigation transfer — not present. No assignment exists and I found no infringement suit naming US 12461238 (granted only 2025-11-04; it is ~10 months post-grant as of today).
6. Bankruptcy fire-sale — not present. No bankruptcy proceedings involving TrinamiX GmbH found.
7. Privateering — not present. No operating-company-to-NPE transfer exists; no SEC filings implicated (TrinamiX is not a US public company).
8. Defensive aggregator (anti-NPE) — not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN; it remains at the originating operating company.

Verdict

Insufficient data — but with an important clarification: this is not a case of missing records on an active NPE chain; it is a case of no recorded assignments at all, which is the signature of a patent that has remained with its original operating-company assignee, TrinamiX GmbH, since filing. The only "signals" available — sole original assignee, active product-development company in the patent's own technical field, accelerating US filing cadence through 2025, and zero conveyances — point away from any NPE pattern. I could not retrieve raw USPTO Assignment Center reel/frame data in my environment, so I cannot certify the absence of a record with 100% certainty; verify at https://assignmentcenter.uspto.gov/ (search "12461238") before relying on this analysis.

I'll search for the specific patent number 12461238 and its cited prior art references.

I found the cited-references list. Let me now retrieve details on the individual references.

I hit my tool-step ceiling before I could verify every reference individually, so I'll give you what the record actually shows, clearly separating verified from unverified, and flagging where the "cited references" list is not the same thing as an anticipation case.

Prior Art Cited on the Face of US 12,461,238 B2

0. Critical framing before the citations

A. What I actually retrieved. The authoritative source for the citation list is the patent's own "Referenced Cited" face data, surfaced at Justia for this exact number:
https://patents.justia.com/patent/12461238#16 → heading "Referenced Cited," with the footer "Patent number: 12461238 … Filed: Jan 30, 2025 … Date of Patent: Nov 4, 2025." I searched the literal number 12461238 and did not substitute similar numbers (e.g., 12,461,235 or 12,150,638); hits on those were discarded.

B. "Cited reference" ≠ "anticipatory prior art." The list below is the union of references cited by the applicant (IDS) and by the examiner, per Google/Justia convention. Several are the applicant's own earlier family members and several are pure background/technology citations that were never relied on for rejection. A § 102 anticipation requires a single reference disclosing every element of a claim, arranged as claimed. On this record, no single cited reference does that for the independent claims as I can reconstruct them.

C. Which claims we are mapping to — and a contradiction I must flag. Building on the earlier sections of this analysis:

• The specification's operative claim construct is the φ1z/φ2m detector (distance-dependent image filter + material-dependent image filter → z and m).
• The Justia-reported granted claims, per the previously generated "Patent summary," are narrower system claims (NIR CMOS sensor + first VCSEL pseudo-random pattern + second light source periodic pattern + DOE + processor comparing an intensity distribution to reference image data → object classification, with dependent claims reciting skin vs. non-living surfaces).

These two constructs do not have the same prior-art exposure, and I flag this as a live contradiction/internal inconsistency in the source record that must be resolved against the official claim text before any § 102 opinion is relied on. I could not retrieve the verbatim granted claim 1.

D. Date qualification matters enormously here, and it cuts both ways. Effective filing date is ~2019-03-15 (the Google Patents "Prior art date"; note the family members US20220157044A1 / EP 3938802 list 2019-03-14 — the previously flagged ~1-day family discrepancy). If the granted claims are entitled to that date, references published after 2019-03-15 are not § 102(a)(1) art; if the granted system claims are not supported by the 2019 priority application (plausible given how different they read), their effective date is 2025-01-30 and every reference below becomes available. I mark each reference accordingly.

1. U.S. patent document citations

2. Foreign patent document citations

3. Non-patent literature citations

4. Bottom line on § 102 anticipation

Building on, rather than repeating, the earlier Obviousness section:

1. No cited reference is a clean § 102 "smoking gun" for the independent claims. The granted independent claims (NIR CMOS + first VCSEL pseudo-random + second periodic light source + DOE + reference-image intensity-distribution comparison → classification) are a specific combination that no single item in the cited list appears to disclose in full. The honest answer to "which claims does X anticipate" for most entries is: the broadest/spec-level claims only, and only if the 2019 priority holds.

2. The three strongest § 102 candidates are citizen-level, not claim-1-level:

   • US 2016/0155006 A1 (Makkapati) → any claim whose "material property" is biological tissue/human skin determined from an imaged projected pattern (directly on point for the granted dependent "skin vs. non-living" claims).
   • US 2018/0033146 A1 (Bleyer) → claims coupling depth + per-dot reflectivity (a material property) from one structured-light image.
   • DE 198 46 619 A1 → the broadest "material-property-from-photosensor-array-signals" claim language.

3. The TrinamiX/BASF self-citations are the closest technical art but the weakest § 102 weapons. Items #3, #5, #13–19 are the applicant's own family. The pre-2018 ones (WO 2012/110924, WO 2014/097181, WO 2014/198629; US 2017/0363465) are true § 102(a)(1) art because they are outside the one-year grace window. But the May-2018 WO 2018/0916xx trio and US 2018/0276843 fall inside one year of 2019-03-15 and name overlapping inventors → a patentee will assert the § 102(b)(1)(A) inventor-disclosure exception, and for any (a)(2) reliance, the § 102(b)(2)(C) common-ownership exception. Expect that fight.

4. Two references are almost certainly not prior art on the spec-level claim construct — US 2019/0378286 (Nash, Dec. 2019) and US 2022/0067410 (Raz, Mar. 2022) both post-date the 2019-03-15 priority date. They appear on the face because prosecution of the 2025 continuation considered them, which is itself evidence that the examiner treated at least some granted claims as having a later effective date than 2019-03-15 — reinforcing the priority-dependency question in ¶ D above. If those claims lack 2019 written-description support, then US 2019/0107389 (Ahmed), US 2019/0378286 (Nash) and US 2022/0067410 (Raz) all become live § 102 art, which materially changes the risk picture.

5. Address the earlier-generated sections' caveat head-on: the prior "Obviousness" analysis treated only the five references discussed in the specification, with US 2018/0033146 + US 2016/0155006 as the primary § 103 combination. That conclusion is unchanged and reinforced — but note the fuller face list adds US 2019/0037136 (Downing, pseudo-random illumination patterns + reflected-radiation detection) and US 2019/0041736 (Grunnet-Jepsen, 3D-camera pattern projector), which are the only cited items that speak directly to the "pseudo-random NIR pattern from a VCSEL/LED projector" limitation the earlier summary attributed to the granted independent claim. If the granted claim really does recite a pseudo-random projected pattern, these two — not the spec's five — are the references to run down first.

5. Explicit verification gaps (do not rely on these without confirming)

• Unverified subject matter: US 2019/0107389 A1 (Ahmed), US 2019/0378286 A1 (Nash), US 2022/0067410 A1 (Raz), and the exact title of US 2019/0041736 A1 (Grunnet-Jepsen). I hit the step limit before pulling full records.
• Unverified verbatim claim text for US 12,461,238 B2 (particularly independent claim 1), so all § 102 mapping is to a reconstructed claim construct.
• Not confirmed whether each listed reference was cited by the examiner (i.e., relied on) or merely by the applicant in an IDS; nor which were cited with a "material/alleged" tag in the family's ISR.
• Assignee/inventor attributions for items 1, 2, 6, 7 are from company knowledge of the named inventors, not from the fetched records.
• The 2019-03-14 vs. 2019-03-15 priority discrepancy (already flagged in the earlier Patent summary) is unresolved and matters for items #16–19 (the May-2018 WO trio edge closer to / further from the grace window depending on which date governs).

Primary sources: Google Patents record for US12461238B2 (authoritative text supplied); https://patents.justia.com/patent/12461238 (Referenced Cited table); https://patents.google.com/patent/US20220157044A1/en (parent family publication); EPO publication-server PDF of EP 3 938 802 B1 (data.epo.org/.../EP3938802NWB1/document.pdf), which reproduces the same background-art characterizations and confirms the application/priority data.

Obviousness Analysis — U.S. Patent No. 12,461,238 B2 (TrinamiX GmbH)

Patent: US12461238B2, "Detector for identifying at least one material property"
Assignee: TrinamiX GmbH | Priority: 2019-03-15 | Grant: 2025-11-04
Status: Active (AIA regime applies; effective filing date post-March 16, 2013, so § 103 as amended by the AIA governs)

Because the supplied patent text does not reproduce the formal claims section, I reconstruct the operative independent claim from the published claim language visible on Justia/Google Patents for this exact patent family (Justia, "Detector for identifying at least one material property," https://patents.justia.com/patent/12461238; and the parent publication https://patents.google.com/patent/US20220157044A1/en). Any claim-element mapping below is keyed to that reconstructed scope and flagged where the granted claim language is not fully confirmed.

1. The claimed invention

The claimed detector for "identifying at least one material property m" requires, in substance:

1. [Sensor] at least one sensor element with a matrix of optical sensors (each with a light-sensitive area), configured to record at least one reflection image of a light beam originating from an object;
2. [Beam-profile evaluation] an evaluation device configured to determine the material property by evaluating at least one beam profile of the reflection image;
3. [Distance feature] the evaluation device determines a distance feature φ1z by applying a distance-dependent image filter φ1 to the reflection image, selected from: a depth-from-photon-ratio (DPR) filter, a depth-from-defocus (DFD) filter, a linear combination, or a further filter correlated to those by |ρ| ≥ 0.40;
4. [Material feature] the evaluation device determines a material feature φ2m by applying a material-dependent image filter φ2 to the reflection image, where the filter "passes a hypothesis testing" (null hypothesis: no distinction between material classifiers; alternative: distinguishes ≥ two; pass if p-value ≤ significance level);
5. [Joint output] the evaluation device determines a longitudinal coordinate z and the material property m by evaluating φ1z and φ2m together.

The specification itself concedes the problem motivating the invention: beam-profile-analysis (BPA) features such as spot width "depend on both, distance and material," making material identification from a reflection profile unreliable.

2. The prior-art record (per the patent's own "Prior Art"/background section)

The specification identifies five pieces of art:

Additional, same-field art is visible in the family listing (e.g., trinamiX's WO 2018/091638 A1, published 2018-05-24, "Detector for optically detecting at least one object," which predates the 2019-03-15 priority date and discloses distance determination from recorded beam profiles), but I treat the five references above as the primary § 103 basis per your instruction.

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

A PHOSITA would hold a degree in physics, electrical engineering, or computer vision (or equivalent experience), with several years working on optical sensors, structured-light/camera systems, image processing, and statistical classification. They would be conversant with: CMOS/IR cameras, dot/line projectors, triangulation and depth-from-defocus, NIR spectroscopy of tissue, and standard image metrics (gradients, GLCM textural statistics, smoothing/speckle-contrast, Law's masks, threshold areas).

4. Proposed § 103 combinations and motivations

Ground A — US 2018/033146 A1 in view of US 2016/155006 A1 (primary combination)

Element mapping:

• Sensor / reflection image: US 2018/033146 teaches a camera capturing a structured-light image containing reflected dots — i.e., a matrix of optical sensors recording a reflection image of projected light returning from an object. US 2016/155006 independently teaches an imaging unit recording a projected IR dot pattern reflected from a scene.
• Beam-profile evaluation for material property: US 2018/033146 derives a reflectivity map from per-dot analysis of the captured image; reflectivity of a surface is itself one of the material properties listed in the '238 specification ("a reflectivity, a specular reflectivity, a diffuse reflectivity"). US 2016/155006 goes further and evaluates per-dot intensity and edge blur (the imaged "beam profile" of each projected feature) caused by wavelength-dependent light diffusion/penetration into skin, expressly to distinguish skin from other materials (cloth, speckle-producing surfaces).
• Distance feature: US 2018/033146 computes depth per dot by triangulation. US 2016/155006 likewise accounts for the distance of the illumination unit when selecting intensity/blur thresholds. DFD is disclosed at least inchoately in US 2016/155006's "edge blur measurement unit," since imaged dot blur varies with both defocus distance and material diffusion — the very ambiguity the '238 patent addresses.
• Material feature: US 2016/155006's comparison of diffusion/absorption factors, edge-blur differences across wavelengths, and intensity differences is a material-dependent evaluation of imaged beam features; the enumerated '238 material filters (spot-shape/translucency correlates, smoothness/speckle-contrast, threshold-area, etc.) are simply alternative image metrics for the same underlying phenomenon (beam broadening and speckle suppression by translucent/scattering materials such as skin).
• Joint z, m determination: US 2018/033146 already determines depth and reflectivity from the same structured-light image; US 2016/155006 already conditions its material/skin decision on the known distance. A PHOSITA combining them would, as a matter of course, evaluate the material-discrimination metric as a function of the measured per-dot distance — i.e., "determining z and m by evaluating the distance feature and the material feature."

Motivation to combine (KSR rationale): Both references solve the same problem — robust identification of material type (skin vs. non-skin) from an actively projected dot pattern imaged by a camera — using the same hardware class (IR dot projector + CMOS imager). US 2018/033146 supplies what US 2016/155006 lacks (a per-dot distance estimate rather than a fixed, known-distance setup), and US 2016/155006 supplies what US 2018/033146 lacks (a material-classification decision rule built on beam-profile/diffusion physics rather than a generic reflectivity map). The combination is the union of two known, compatible modules in the same field, yielding the predictable result that material features are distance-normalized before classification. This is the classic KSR "combination of familiar elements according to known methods" that does "no more than yield predictable results."

Ground B — US 2016/155006 A1 + CN 108 363 482 A (or + US 2018/033146 A1) for the gesture/interface embodiments

CN 108 363 482 teaches a binocular structured-light system that reconstructs 3-D geometry of a user and segments body parts for gesture control. The '238 claims contemplate identical downstream uses (human-machine interface, entertainment device, tracking). Motivation: any interactive system that segments a human hand from a 3-D reconstruction must first know which reconstructed pixels are skin — the express problem of US 2016/155006. A PHOSITA designing a gesture-controlled TV or phone (the '238 specification's own stated field) would combine the 3-D/structured-light pipeline (CN 108 363 482; US 2018/033146) with the skin-discrimination-from-beam-profile analysis (US 2016/155006) because the combined system provides both the depth (needed to know where the object is and to normalize the pattern) and the material decision (needed to know what it is). This directly supplies the claimed "determine z and m by evaluating φ1z and φ2m."

Ground C — US 2018/033146 A1 + DE 198 46 619 A1

DE 198 46 619 teaches deriving a structure/material code from photosensor-array signals — i.e., an evaluation unit that converts a recorded optical reflection pattern into a material/surface classifier. US 2018/033146 teaches the depth-normalized, per-dot reflectivity analysis of a structured-light image. Motivation: DE's structure code is ambiguous unless the measurement geometry (distance) is known; US 2018/033146 supplies per-dot depth from the same captured dots; combining the two yields distance-robust material classification from the beam profile — the precise technical effect claimed. Both references address "determining a characteristic of a surface from the electrical output of an imaging sensor array," making the combination a routine aggregation with predictable benefit.

Ground D — US 2016/0206216 A1 as a supporting secondary reference

US 2016/0206216 teaches multi-signal skin detection (thermal + optical) with an evaluation unit fusing sensor channels to detect skin. It evidences that PHOSITAs in this exact field (trinamiX skin-detection work) routinely (i) build skin/material classifiers on optical reflection data and (ii) fuse a second, independent measurement channel to disambiguate skin from look-alike materials. Used with Grounds A–C, it supplies the motivation to add an independent channel/feature (here: a distance feature) to make the material feature reliable — reinforcing the reason to combine depth estimation with material-feature extraction in one evaluation device.

Ground E — Known distance-from-blur and distance-from-photon-ratio techniques (supporting)

The "distance-dependent image filter" limitation is the claim element most generic on its face: DFD is a decades-old technique, and DPR (photon-ratio depth from defocused beam profiles) was published by the same art community (e.g., trinamiX's pre-2019 publications such as WO 2018/091638 A1 and the WO 2014/097181 A1 family, the latter being cited in the '238 specification itself for transversal sensors and related optical distance detection). A PHOSITA implementing the Ground A–C systems would select one of these known distance estimators essentially by design choice. The "further filter correlated |ρ| ≥ 0.40" fallback is a mathematical restatement ("any filter whose output tracks a distance-dependent filter"), and the correlation/similarity test (Pearson's ρ) is a conventional statistical tool.

5. Graham-factor summary

1. Scope and content of prior art: Each claim element maps onto the art: reflection imaging of projected patterns (all five refs.), beam-profile/pattern evaluation for material or surface classification (US 2016/155006; DE 198 46 619; US 2018/033146), distance measurement from the same image (CN 108 363 482; US 2018/033146; DFD/DPR knowledge), and joint distance + material analysis (US 2018/033146's depth-plus-reflectivity; US 2016/155006's distance-conditioned segmentation).
2. Differences: The principal differences are (i) the formal separation of the evaluation into a distance-filter branch and a material-filter branch, (ii) the statistical selection rule ("passes a hypothesis testing," p ≤ significance level, e.g., p ≤ 0.075–0.01) for choosing the material filter, and (iii) correlation-threshold language (|ρ| ≥ 0.40–0.80).
3. Level of ordinary skill: As above — a vision/optics engineer familiar with structured light, image statistics, and material/skin spectroscopy.
4. Secondary considerations: None are evident from the record provided (no unexpected-results showing, no long-felt-need evidence, no commercial-success data). The specification itself frames the advance as using "known" image filters (luminance, spot-shape, gradient, standard deviation, smoothing, GLCM, Law's energy, threshold-area — all textbook operators) and known depth techniques (DPR, DFD) in a combined evaluation — language that undercuts any claim of a non-obvious component breakthrough.

6. Where the combination is strongest and where it is vulnerable

Strongest § 103 case: Element combination [sensor] + [beam-profile evaluation] + [distance feature] + [joint z/m output] is squarely taught by US 2018/033146 (depth + reflectivity from one dot image) read with US 2016/155006 (material/skin decision from imaged-dot blur/intensity). The '238 specification's own admission — that BPA features mix distance and material information and that this mixing is the problem — actually helps an obviousness challenger: the natural, well-trodden fix (used in both primary references) is to measure distance separately and condition the material metric on it. Under KSR, where "the prior art … provides a reason to combine," and where there is a "finite number of identified, predictable solutions," the combined system would have been obvious.

Vulnerable/plausible patentee counterarguments:

• The hypothesis-testing limitation. No listed reference discloses applying ANOVA/F-testing (or any statistical hypothesis test) to select the material filter against material classifiers. A patentee could argue this is a specific, non-obvious design methodology. However, the underlying material filters are conceded to be known; using a standard significance test to choose among known candidate filters is a routine statistical-optimization step that examiners and courts frequently treat as obvious (an "obvious to try" with a reasonable expectation of success, or mere optimization/design choice). The p-value thresholds (≤0.075/0.05/…) are conventional significance levels.
• Correlation thresholds (|ρ| ≥ 0.40, 0.60, 0.80). These are arbitrary numerical design choices applied to a known similarity metric (Pearson's ρ); dependent-claim numerology of this kind rarely withstands § 103 where the underlying filter families are known.
• Specific material filters. If the material feature is tied to a particular filter (e.g., a speckle-contrast/smoothness filter computed as std-dev of a despeckled ratio, or Law's energy filters), the art still shows the underlying physical effect (US 2016/155006 discusses speckle on non-skin objects and diffusion blur on skin); implementing it with any of the named standard image operators would have been within the PHOSITA's routine toolkit.

Overall assessment: Under § 103 and Graham/KSR, the central claim scope — an optical detector that evaluates a reflected illumination pattern both for distance (via DPR/DFD/triangulation-type filtering) and for a material-discriminating image feature, and that reports z and the material property from that joint evaluation — is vulnerable to obviousness over US 2018/033146 A1 combined with US 2016/155006 A1 (with CN 108 363 482 A and DE 198 46 619 A1 available to supply the structured-light depth and sensor-array material-code teachings, respectively, in alternative grounds). The patent's most defensible ground is the narrow, method-level claim language requiring the specific statistical hypothesis-testing filter-selection protocol, which does not appear verbatim in the five references discussed in the specification; whether that limitation is enough to preserve validity would turn on claim construction and on whether the challenger can show the statistical-selection step was itself conventional in image-feature engineering.

Sources used

• Patent text as supplied (https://patents.google.com/patent/[US12461238](/patent/US12461238)/en; fetched 2026-09-04).
• Justia claim/description excerpt for US12461238: https://patents.justia.com/patent/12461238
• Family publication US20220157044A1: https://patents.google.com/patent/US20220157044A1/en
• US 2016/155006 A1 ("Device and method for skin detection") technical detail: https://www.patentsencyclopedia.com/app/20160155006

Caveat: the granted claim set of US12461238B2 was not fully reproduced in the supplied text; element-level mapping above is based on the reconstructed independent claim and should be re-verified against the official claim text before reliance in any proceeding.