US 11896346

I'll research this patent number directly.

Let me dig into the litigation and post-grant proceedings involving this specific patent.

Let me check for Federal Circuit 2026 activity and the specific asserted-claim details for the '346 patent.

US 11,896,346 B2 — Analyst Summary

Search scope note: I searched for the exact identifier 11896346. Results returned only US 11,896,346 B2 and its own pre‑grant publication (US 2023/0172455 A1). No other patent number matched. One search returned a UK Companies House entry containing the unrelated company‑registration number "11896346" — that is a coincidental numeric match and is excluded.

---

1. Bibliographic data

Abstract. The fetched Google Patents record did not surface the verbatim abstract paragraph, and my searches did not return it. I therefore will not quote it. The disclosure summarises the invention as an optical system in which a light source of semiconductor diodes (LEDs and/or laser diodes, optionally with Bragg reflectors) emits near‑IR/SWIR light, is delivered through a housing/encapsulating layer (enamel, dentine, plastics, packaging) to tissue or an object, and reflected light is detected by synchronized photo‑detectors (optionally InGaAs, optionally an image‑sensor array with CMOS) after passing wavelength‑selective dielectric filters; a processor identifies the object/user based on water absorption by comparing returns at a lower‑water‑absorption wavelength (near 1090 nm) against a higher‑water‑absorption wavelength (near 1440 nm) and comparing the resulting value to a threshold. Treat this as a paraphrase of the disclosure, not the abstract of record.

---

2. The patent's core technical idea

The disclosed invention is an optical identification/sensing system positioned around the SWIR water‑absorption bands. The specification notes that water has transmission windows at ≈1500–1850 nm and ≈2050–2500 nm (Fig. 3A), that water absorption is a discriminating signal, and that 1400–2500 nm light sits in an eye‑safe wavelength window. The claimed architecture therefore pairs (a) a low‑water‑absorption channel (~1090 nm) with (b) a high‑water‑absorption channel (~1440 nm), modulates the sources, and uses synchronized detection (with the detector optionally able to sample while the emitters are off, in the broader family) to produce a reflectance ratio that is compared against a threshold. Notably, the claims are written to cover not just skin, but also teeth and packaging/encapsulating layers, consistent with the specification's caries‑detection and through‑package pharmaceutical screening discussion (Goog. Patents description, e.g., A61B5/0088 and G01N21/359 classifications).

---

3. Independent claims in plain language

Based on the Justia claim text for US 11,896,346 and the claim characterisation in the 2026 W.D. Tex. complaint (Omni MedSci, Inc. v. Apple Inc., 7:26‑cv‑00226, ¶¶ 79–81), the granted patent has an independent claim 1 and an independent claim 6, with the remaining claims dependent.

Independent claim 1 — "sensor system" that identifies a user by water content. An active illuminator made of several semiconductor diodes (modulated), each emitting at one or more optical wavelengths in the 700–2500 nm near‑IR range. At least one diode emits a first wavelength with lower water absorption, near 1090 nm; at least a second emits a second wavelength with higher water absorption, near 1440 nm. One or more wavelength‑selective optical filters (dielectric filters) sit in front of the illuminator. A housing takes the filtered diode light and delivers it to the user, and is coupled to an electrical circuit and a processor. A detection apparatus with one or more photo‑detectors receives reflected diode light and is synchronized to the illuminator. The processor identifies the user based on water absorption within the user by generating a first signal part from the 1090‑nm return, a second signal part from the 1440‑nm return, and comparing them to produce an output value.

Independent claim 6 — optical system that identifies an object by water content, with a threshold test. A light source of several modulated semiconductor sources emitting at least partly in the 700–2500 nm near‑IR range, again with a first modulated LED near 1090 nm (lower water absorption) and a second modulated LED near 1440 nm (higher water absorption). One or more wavelength‑selective optical filters (dielectric) are placed in front of the light source to pass at least part of the first or second wavelength. A housing receives the passed light, delivers an output to an object, and is coupled to an electrical circuit and a processor. A detection system with one or more photo‑detectors receives reflected output and is synchronized to the first or second modulated LED. The processor identifies the object based on water absorption by generating first/second signal parts from the two wavelengths and comparing them to generate an output value, and then compares the output value to a threshold.

Third independent-claim concept (uncertain). The Specification's summary language also recites an optical system for detecting a tissue or object using semiconductor diodes emitting a SWIR wavelength between 1400 and 2500 nm corresponding to an eye‑safe window, delivering through an encapsulating layer, with a detection system synchronized to the light source and a dielectric detection‑side filter between the object and the detector. In the pre‑grant publication (US 2023/0172455 A1) this was written as claim 6, while the 1090/1440 claim was written as claim 15. In the granted patent, claim 6 is the 1090/1440 claim, which indicates the claim set was renumbered/amended during prosecution. Whether the eye‑safe‑window claim survives in the granted claim set, the total claim count, and the exact dependency map beyond claims 1–11 are things I could not verify with authority; treat the numbering beyond claims 1 and 6 as uncertain.

Representative dependent claims (from the Justia text for the '346 patent):

• Detector‑side wavelength‑selective dielectric filter in addition to the source‑side filter.
• Explicit 1090 nm / 1440 nm recitation.
• Delivery to skin, tissue or teeth of the user, plus the threshold comparison.
• Modulated LEDs; photo‑detectors comprising indium gallium arsenide (InGaAs).
• Image sensor comprising an array of photo‑detectors coupled to CMOS transistors.
• Encapsulating layer = enamel, dentine, plastics or packaging material.
• Sources = LEDs, or laser diodes comprising one or more Bragg reflectors.
• Optical polarizer between source and detector to improve SNR.
• Pattern‑matching technique to identify the user's tissue.
• Coupling to a wearable device, and coupling to a smart phone or tablet; pattern recognition; remote‑sensing/time‑of‑flight variants (complaint describes claims 7–11 along these lines).

---

4. Litigation and PTAB posture relevant to this exact patent

Note on dates: my instructions state today is 26 April 2026 (the system clock shows 28 September 2026). One source I retrieved reports a complaint filed 8 June 2026 — i.e., after the stated "current" date. I am reporting the document literally rather than normalising the date.

• Omni MedSci, Inc. v. Apple Inc., No. 7:26‑cv‑00226 (W.D. Tex.). Complaint (Doc. 1) asserting, among others, US 11,896,346. The complaint expressly analyses the '346 Patent's independent claim 1 (user identification via 1090/1440 nm water absorption; dependents 3, 4, 5) and independent claim 6 (object identification via 1090/1440 nm modulated sources, dielectric filters, housing, synchronized detection, threshold), with claims 7–11 as further enhancements. It also states the '475 patent (US 12,193,790), US 11,564,577 and the '346 patent are related through a chain of continuations with a 31 December 2012 priority. Source: CourtListener docket 73456916.
• Earlier Apple litigation. Omni MedSci v. Apple, E.D. Tex. 2:18‑cv‑00134; N.D. Cal. 4:19‑cv‑05924, 4:20‑cv‑00563 — these asserted the '533, '546, '286, '698, '484 families, not the '346 patent.
• Samsung/Fossil/Oura/OnePlus (E.D. Tex. 2:24‑cv‑01070‑JRG‑RSP) and WHOOP (D. Del. 1:25‑cv‑00140‑JLH). These attacks target US 9,055,868; 9,651,533; 10,874,304; 11,160,455; 12,193,790; 12,268,475 — i.e., not 11,896,346. IPR2025‑01249 through ‑01254 and PGR2025‑00063/00064.
  • US 11,896,346 does appear as an exhibit in those proceedings: "Petitioner Exhibit 1088: Ex. 1088 – US11896346" in PGR2025‑00064 (Samsung v. Omni MedSci, '790 patent), and the '790 file history shows a terminal disclaimer filed over several family patents "including U.S. Patent No. 11,896,346." In other words, the '346 is cited as a related/priority member, not as the challenged patent.
• I found no IPR, PGR, or EPR petition whose challenged patent is 11,896,346.
• Federal Circuit. I found no 2026 CAFC docket naming US 11,896,346. The live Omni/Apple appellate matter is Omni MedSci, Inc. v. Apple Inc., No. 25‑1646 (Fed. Cir.), which arises from the '484 IPR2021‑00453 remand decision (Board held claims 3–6 and 8–14 unpatentable on remand, 14 Feb 2025) and does not involve the '346. Earlier CAFC decisions in this family: Nos. 20‑1715, 21‑1229 (summary affirmance of '533 IPR, 8 June 2022), 23‑1034 (Apple v. Omni MedSci, 21 June 2024, aff‑in‑part/vac‑in‑part/remand on the "identify an object" construction).

---

5. Explicit uncertainties

1. Verbatim abstract — not retrieved; I paraphrased and said so.
2. Total claim count and full dependency structure of the granted '346 — I directly verified the independent claim 1 and claim 6 texts and dependent-claim subject matter, but not the complete set (the pre‑grant publication had a differently numbered three‑independent structure).
3. Asserted claims in the 2026 Apple case — the complaint discusses claims 1 and 6 and their dependents; I did not retrieve an itemised assert list.
4. Assignee naming — "Omni Medsci, LLC" (USPTO assignment record per Google Patents) vs. "Omni MedSci, Inc." (litigation). I have not reconciled these.
5. Exhibit 1088's purpose in PGR2025‑00064 — appeared as a cited exhibit; I did not confirm whether it was used as prior art or as evidence of the family/priority chain.
6. Any statement above that is not tied to the Google Patents record or to the cited docket/PTAB documents should be treated as my inference, not authoritative.

I'll search for litigation involving US patent 11896346.

Let me search more specifically using litigation databases and the patent owner's name.

Let me search specifically for the '346 patent number in litigation contexts.

Bottom line

Based on the searches I was able to run (general web, CourtListener/Justia/Docket Alarm/Unified Patents/law‑firm alerts), I could not verify any lawsuit in which U.S. Patent No. 11,896,346 ("the '346 patent") is specifically pleaded as an asserted patent‑in‑suit. I want to flag that clearly rather than pad out a false list, because the '346 is easy to confuse with its many Omni MedSci sibling patents.

What I can confirm is:

1. Ownership/identity. US 11,896,346 B2, "Short‑wave infrared sensor for identifying based on water content," inventor Mohammed N. Islam, application 18/103,408, filed Jan. 30, 2023, granted Feb. 13, 2024, priority date 2012‑12‑31, current assignee Omni MedSci, LLC (originally Omni MedSci, Inc.). It is a member of the same continuation family as U.S. Pat. Nos. 11,564,577; 11,896,346; 12,193,790; 12,268,475; 12,484,787 and the earlier '868, '533, '304, '484, '455 patents.
2. The '346 does appear in a PTAB record — as an exhibit, not as the challenged patent. In Samsung Electronics Co., Ltd. et al. v. Omni MedSci, Inc., PGR2025‑00064 (P.T.A.B.) (filed Aug. 5, 2025; terminated Dec. 1, 2025 by discretionary denial), Petitioner Exhibit 1088 is listed as "US11896346." So the '346 was pulled into that post‑grant‑review record as background/related‑family material. It was not the patent under review (PGR2025‑00064 addressed U.S. 12,193,790).
3. Google Patents' family page carries a litigation flag. The '346's Google Patents record shows "Family has litigation" with a Darts‑IP family link (family=69228103). That is a family‑level indicator, and I could not resolve it to a case naming the '346 itself. (Verify via Darts‑IP, Unified Patents' litigation portal, or PACER before relying on it — cite: https://patents.google.com/patent/US11896346/en )

Related Omni MedSci litigation (family context — the '346 is NOT the pleaded patent in these)

These are the matters in the same patent family, where sibling patents (not the '346) are asserted. I'm listing them because they're the surrounding landscape and because the '346 may be added to amended complaints.

Related PTAB proceedings (post‑grant/invalidity track)

• Apple v. Omni MedSci, IPR2019‑00916 ('533) — FWD Oct. 14, 2020: challenged claims unpatentable; affirmed, Omni MedSci v. Apple, No. 21‑1229 (Fed. Cir. June 8, 2022).
• Apple v. Omni MedSci, IPR2021‑00453 ('484) — FWD Aug. 3, 2022; Fed. Cir. reversed/remanded (No. 2023‑1034, June 21, 2024); remand FWD Feb. 14, 2025 held claims 3–6, 8–14 unpatentable; Omni appeal pending (Fed. Cir. No. 25‑1646).
• Apple v. Omni MedSci, IPR2020‑00029 ('546) — instituted.
• Samsung (with Fossil and others) v. Omni MedSci: IPR2025‑01249 ('868), IPR2025‑01250 ('533), IPR2025‑01251 ('304), IPR2025‑01252 ('455), IPR2025‑01254 ('475), and PGR2025‑00064 ('790 — the proceeding in which Ex. 1088 = US11896346 appears), filed Aug. 5, 2025, terminated Dec. 1, 2025 by discretionary denial.

Important caveats

• Do not treat the '346 as a litigated patent without further verification. The only docket hit I found tying the exact number 11,896,346 to a proceeding is its use as an exhibit in PGR2025‑00064; that is not an infringement assertion.
• A related AIA/effective‑filing challenge (Director Squires) surfaced for an Omni MedSci wearable patent (post‑AIA cut‑off issue) — I could not tie that specifically to the '346.
• Note that some of the search results are dated after your stated current date (April 26, 2026) — e.g., the June 8, 2026 W.D. Tex. Apple complaint (7:26‑cv‑00226) and August 2026 PTAB exhibits. I'm reporting them as returned; confirm the live docket before relying on them.
• Search limitations: I hit the tool's step limit before running direct queries on PACER/Docket Alarm/Unified Patents for "11,896,346." To close the gap, the highest‑value next checks are: (a) Unified Patents' litigation caselist, (b) Darts‑IP family 69228103, (c) PACER docket searches for "11,896,346" in the E.D. Tex./W.D. Tex. Omni MedSci cases above, and (d) the live complaints in 2:24‑cv‑01070 and the Whoop case to see whether the '346 was added by amendment.

If you'd like, I can draft the specific PACER/Docket Alarm query strings for those four checks, or pull the District Court docket sheets and exhibit lists once search access is restored.

I'll verify the structured-data default with targeted web searches before writing up.

Let me run a couple more targeted checks before finalizing.

Two possible discrepancies surfaced. Let me check them.

Proceedings overview

AIA trial proceedings on file for U.S. Patent No. 11,896,346: 0. The USPTO Open Data Portal (ODP) returns no AIA trials in which the '346 is the challenged patent — 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials. My independent web checks (PTAB docket aggregators, Board papers from the Omni MedSci family, and Federal Circuit records) surfaced no petition for IPR, PGR, or CBM naming the '346 as the patent under review; every hit tying the number 11,896,346 to a PTAB record is an exhibit or family-reference citation inside a proceeding on a different Omni patent, not a challenge to this patent.

Bottom-line defensive posture: this is not a "hardened" patent and it is not a "dead claims" patent — it is an untested one. No claim of the '346 has ever been construed, instituted against, amended, canceled, or confirmed by the Board. That cuts both ways for a defendant: you have no free kill-shot from a prior FWD, but you also have no adverse § 315(e)(2) estoppel, no Board precedent on claim scope, and (importantly) the '346 issued 2024-02-13, so it has been in force only ~2.5 years — recent issuance and first assertion are the classic pre-conditions for a fresh petition, not evidence of resilience.

⚠️ Practical caveat up front: because the '346 appears to have just entered litigation (see discrepancy flag below), the "no PTAB activity" default is time-sensitive. The absence of a petition today is unsurprising and signals little; the presence or absence of a petition 12 months from service will signal a great deal.

---

Proceedings naming US 11,896,346 as the challenged patent

NONE ON FILE

• Type: n/a
• Filed: n/a
• Status: No AIA trial proceeding on file (USPTO ODP), as of the most recent ingest.
• Judge panel: n/a
• Petition grounds: n/a — no petition has ever been filed against this patent.
• Institution decision: n/a
• Final Written Decision: n/a — no claim of the '346 has been canceled, held unpatentable, or upheld by the Board. Any statement to the contrary would be fabrication.
• Settlement / termination: n/a
• Appeal: n/a — no FWD exists to appeal, so there is no CAFC appeal of a '346 trial decision.
• Defensive value: You cannot rely on a prior Board invalidation, and the patent owner cannot point to a Board affirmance. All of claims 1–11 remain presumptively valid under § 282 and are in play both in IPR and in district court. Conversely, no petitioner-side estoppel exists — you would be a first mover on the merits.

Why the default is credible here (and what would change it)

Three structural facts, all grounded in the family's public record, explain the empty docket:

1. The '346 issued late (2024-02-13). It is a large continuation family; Omni has been serializing claims across new continuations, so each new patent restarts the clock as a fresh target. Family members that do have proceedings are the older ones ('533, '304, '455, '868, '790, '475).
2. PGR is almost certainly unavailable. The '346 claims priority to 2012-12-31, making it a pre-AIA patent. Samsung expressly "did not contest" Omni's position that sibling patent 12,193,790 "has a priority date of December 31, 2012 and thus is not an [AIA] patent." See Director's sua sponte review order, IPR2025-01253 (Paper 20). Even if a petitioner argued PGR eligibility, the § 321(c) nine-month window on the '346 closed on or about 2024-11-13, roughly nine months after grant.
   • Nuance worth knowing: the family's AIA-eligibility is live and contested. WHOOP argued the '790 was PGR-eligible and successfully obtained institution in PGR2026-00003 (institution granted 2026-04-01), which prompted the Director to sua sponte review IPR eligibility in the parallel IPR2025-01253. So "pre-AIA, therefore no PGR" is the right default for the '346, but at least one panel/Director has entertained the opposite argument on a same-priority sibling. In any event the § 321(c) window has lapsed, so IPR is the only realistic AIA vehicle for the '346.
3. The '346 has not been a high-value target until now — its siblings carried the litigation load in the 2018–2025 Apple/Samsung/WHOOP campaigns.

---

Where the '346 does appear in the PTAB record — evidence, not challenge

These are not proceedings on the '346. They matter only as (a) proof the family is being worked over, and (b) potential prosecution-history/admission material.

---

Strategic summary

Claim status of the '346. Every claim is UNTESTED. Nothing is CANCELED; nothing is SUSTAINED. The published record identifies the asserted claim set in the new litigation as independent claims 1 and 6 and dependent claims 2–5 and 7–11 (per the complaint analysis of W.D. Tex. 7:26-cv-00226; see discrepancy flag). If you need a defensive narrative today, it has to be built from (i) the Board's FWDs on related patents ('533 all challenged claims unpatentable, aff'd, Omni MedSci v. Apple, No. 21-1229 (Fed. Cir. 2022-06-08); '484 claims 3–6 and 8–14 held unpatentable on remand, FWD 2025-02-14, on appeal Fed. Cir. No. 25-1646) and (ii) the district-court record — not from any holding about the '346 itself. Samsung's own petition language ("the vast majority of the limitations challenged in the IPRs… are identical or substantially similar to those already found unpatentable in prior IPRs challenging related patents") is the single most useful admission in the file for a defendant; it is a roadmap, not a judgment.

Estoppel landscape. No § 315(e)(2) estoppel runs against anyone as to the '346, because no FWD has issued in a proceeding on this patent. Two refinements:

• Estoppel on the sibling patents ('533, '484, '790, '475, '868, '304, '455) does not attach to the '346 — a different patent is a different "patent" for § 315(e)(2) purposes. So Apple, Samsung, WHOOP and their privies remain free to petition against the '346 (subject to the one-year bar and to discretionary factors).
• The real constraints a new petitioner faces are discretionary, not statutory: General Plastic (follow-on petitions by the same family of challengers), § 325(d) (the Office has now seen the same family art repeatedly), and Director Squires' institution regime. Note the Oct. 2025 proposed rules (85 FR / Willkie summary, PTAB Announces New Institution Policy) which would (a) require a petitioner to abandon all § 102/§ 103 defenses in every other forum to obtain IPR, and (b) make denial mandatory where a claim was previously upheld by the USPTO/ITC/court. Those are proposals, not effective rules — verify current status before relying on them. They would apply only to IPR, not PGR (already unavailable here).

Pattern signals. High-intensity, serial family litigation on the petitioner side: Apple filed a wave of IPRs (IPR2019-00910 through -00916, IPR2021-00453); Samsung (with Fossil, Fossil Stores I, Fossil Partners, Oura Health, OnePlus) filed IPR2025-01249–01254 plus PGR2025-00063/-00064 on six patents; WHOOP filed its own IPR2025-01583 ('533), IPR2025-01584 ('304), IPR2025-01585 ('455) and PGR2026-00003 ('790). So multiple, independent, repeat players have built the invalidity infrastructure — but none has aimed it at the '346. The patent owner (Omni MedSci) appeals aggressively (Nos. 21-1229, 20-1715, 23-1034, 25-1646), so expect any adverse '346 FWD to be appealed. I found no defensive aggregator (e.g., Unified Patents, RPX) in the chain on the '346 or its siblings.

---

Recommended next steps

1. Start the clock. The '346 appears to have been asserted in Omni MedSci, Inc. v. Apple Inc., W.D. Tex. 7:26-cv-00226, complaint filed 2026-06-08, with summons returned executed on Apple 2026-06-10. Under 35 U.S.C. § 315(b), Apple (and its RPI/privies) has until on or about 2027-06-10 to petition for IPR on the '346. [Docket: https://www.courtlistener.com/docket/73456916/omni-medsci-inc-v-apple-inc/]
2. Ignore PGR. The § 321(c) nine-month window closed on or about 2024-11-13 (grant 2024-02-13), and the patent is pre-AIA on its face (priority 2012-12-31). Build the offensive case as an IPR under §§ 102/103 only. § 112 and § 101 theories cannot be raised in IPR — they must be litigated in district court/ITC (or, historically, PGR, which is closed).
3. Reuse Samsung's work product. Samsung's IPR/PGR petitions and exhibit lists (the '533 FWD, the '484 FWD and remand FWD, the CAFC opinions at Nos. 21-1229 and 23-1034) are public and directly usable as the starting evidence set for a '346 petition built on the "nearly identical claim limitations" theory. Pull Exs. 1005–1017 from IPR2025-01250/-01253 and Ex. 1088 from PGR2025-00064.
4. Consider ex parte reexamination as a complement. Given post-2025 institution-rate compression and the record 607 procedural denials in 2025 (Unified Patents 2025 in Review), an ex parte reexam request — no standing, no time bar, no § 315(b) clock, and (>90%) grant rates historically — is a low-cost parallel track to create a § 325(d)/prior-art record even if an IPR is discretionarily denied.
5. Verify the docket state mechanically. Run direct queries against (a) PTAB E2E/PTACTS for the '346 application number 18/103,408 and patent number 11,896,346; (b) patentcenter.USPTO.gov for any newly filed IPR/PGR; (c) the ODP AIA trial API. Flag immediately if a petition appears — the ODP ingest lags several days to weeks.

---

Discrepancy flag (contradicts an earlier section of this analysis)

The previously generated Litigation summary states that the 2026-06-08 W.D. Tex. Apple complaint "asserts '475 and '787" and "expressly identifies 11,896,346 as a family relative of '475 (not itself asserted)." That is contradicted by two independent sources I retrieved:

• The complaint text itself references "the asserted claims of the '346 Patent" (Compl. ¶¶ 80–81, at page 20 of 65), [https://www.courtlistener.com/docket/73456916/1/omni-medsci-inc-v-apple-inc/]; and
• Third-party complaint analysis for 7:26-cv-00226 lists U.S. Patent No. 11,896,346 as an asserted patent with "Independent claims 1, 6, and dependent claims 2–5, 7–11," accused against Apple AirPods with "heart rate sensors and skin sensors" (¶166). [https://ai-lab.exparte.com/case/dct/txwd/7:26-cv-00226/doc/analysis/1]

If the '346 is in fact asserted, that changes the analysis materially — it starts the § 315(b) clock (2027-06-10), makes the "no PTAB activity" observation time-sensitive rather than dispositive, and means the patent owner's "untested" posture is likely to end within the year. Confirm by pulling the full complaint (Docket 1) and its exhibits A–N before relying on either version. I could not retrieve the full complaint text, and my attempts to search for a 2026 IPR petition specifically targeting the '346 were cut off by the tool's step limit — so treat "zero proceedings" as verified for the ODP ingest and the searches I ran, not as a guarantee against a very recently filed petition that has not yet been indexed.

I'll research the assignment record for US 11,896,346. Let me start with targeted searches.

Let me search for the specific assignment reel/frame and Omni MedSci's corporate history.

Let me dig into the specific assignment records and corporate entity details.

Assignment Record Reconstruction — US 11,896,346 ("Short-wave infrared sensor for identifying based on water content")

Retrieval status / honesty flag up front: I was able to run only a limited number of web queries before hitting the tool step-limit, and none of them returned the USPTO Assignment Center record itself (i.e., the reel/frame number, conveyance-type field, or the correspondent-of-record field on the assignment document). What follows is grounded in (a) the patent's own Google Patents legal-events feed, (b) Federal Circuit / PTAB records that describe the family's 2013 assignment, and (c) prosecution file histories that expose the attorney of record. I have not fabricated reel/frame numbers or correspondent names for the assignment documents. Where a field is unverified I say so explicitly. Verify at https://assignment.uspto.gov/patent/index.html (search "11896346") or https://assignmentcenter.uspto.gov/.

---

Inventors

Unusual pattern flagged — a litigated cloud on title, not a multi-inventor departure. This is a single-inventor family, so the usual "all inventors left within 12 months" tell does not apply. The distinctive anomaly is that the sole inventor's university employer asserted an ownership claim to this family and litigated it: UM's Office of Technology Transfer refused to release UM's rights (May 2, 2013 determination), and UM later filed an amicus brief asserting ownership in Omni MedSci, Inc. v. Apple Inc., 7 F.4th 1148 (Fed. Cir. Aug. 2, 2021) (Linn, J.; Newman, J., dissenting). The Federal Circuit held Bylaw 3.10's "shall be the property of the University" language was not a present automatic assignment, which saved Omni's standing. That decision is the single most important fact for chain-of-title analysis of every patent in this family, including the '346.

---

Original assignee

Omni MedSci, LLC — this is what the Google Patents record lists as both original and current assignee of US 11,896,346. Note the entity-name inconsistency across the family: the earlier sibling patents (e.g., the '533, '304, '455, '546) print as "OMNI MEDSCI, INC., Ann Arbor, MI (US)" on their faces, while the '346's record carries Omni MedSci, LLC. The 2023-09-21 reassignment event changes the assignee from (per the record text) Mohammed N. Islam to OMNI MEDSCI, LLC. I could not retrieve the recorded document to confirm whether this was a direct inventor→LLC conveyance or a correction/confirmatory instrument tied to an Inc.→LLC conversion.

• Primary line of business: fiber-laser / short-wave-infrared light-source development and non-invasive optical physiological sensing — specifically a wearable, non-invasive glucose monitor. Founded by Dr. Islam; Crain's Detroit Business lists Omni MedSci among Michigan's "Most Innovative Companies" describing it as making "a wearable glucose monitor that tracks blood sugar — without drawing blood."
• Did it ship a product embodying the claims? Not verified. I found no evidence of a commercially available Omni MedSci product, no regulatory clearance, and no sales/revenue disclosure (Omni MedSci is private; no SEC filings). Descriptions of the product are aspirational/press-profile. The observable revenue-generating activity is patent assertion.
• Current status: Operating as a private, inventor-controlled licensing/enforcement entity. No bankruptcy, no acquisition, no dissolution found.

---

Assignment timeline

Caveat: The USPTO Assignment Center fields (reel/frame, execution vs. recording date, conveyance type as recorded, and correspondent of record) were not retrievable with the search access available. The two entries below are documented from Google Patents legal events and the Federal Circuit record; a §290 court-action notice and the prosecution correspondence are appended for completeness.

• 2013-12-17 (executed) / recorded on or about 2013-12-17 — Reel/frame: NOT RETRIEVED

  • Conveyance: Assignment (per Omni MedSci v. Apple, 7 F.4th 1148, and contemporaneous firm commentary: "On December 17, 2013, Dr. Islam assigned his patent rights to Omni MedSci; the assignments were recorded at the USPTO.")
  • Assignor: Mohammed N. Islam
  • Assignee: Omni MedSci, Inc. (Ann Arbor, MI)
  • Correspondent: NOT RETRIEVED on the assignment record. (The portfolio's prosecution correspondent is Brooks Kushman P.C./Cheetah Omni MedSci — see below.)
  • Context: Founding transfer — inventor assigned the priority family (provisionals filed 2012-12-31) to his own startup. Whether this instrument expressly sweeps in the later-filed '346 depends on its standard "all continuations/divisionals" language, which I could not read.

• 2023-09-21 (recorded) — Reel/frame: NOT RETRIEVED

  • Conveyance: Assignment of assignors' interest (Google Patents legal-events label: "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: ISLAM, MOHAMMED N.
  • Assignee: OMNI MEDSCI, LLC
  • Correspondent: NOT RETRIEVED (see prosecution-correspondent note under Signal 3)
  • Context: Confirmatory / nunc pro tunc clean-up of title for the '346, recorded ~8 months after the 2023-01-30 filing of application 18/103,408 and ~5 months before the 2024-02-13 grant — most plausibly to perfect the record in light of the litigated 2021 standing decision. Reported to Omni MedSci, LLC; recorded Sep 21, 2023; the assignee and inventor are the same controlling party, so this is not an arm's-length transfer. (Inference on characterization — the underlying document was not retrieved.)

• 2023-09-21 — Reel/frame: NOT RETRIEVED — Change of name / entity-conversion event (unconfirmed). I could not confirm whether a separate Inc.→LLC conversion instrument exists. Treat as unverified.

Non-assignment chain-of-title events (for context, not §Assignment-Center links):

• 2018-04-06 — §290 notice of suit: Omni MedSci, Inc. v. Apple Inc., E.D. Tex. 2:18-cv-00134-RWS, listing patents 9,651,533; 9,757,040; 9,861,286; 9,885,698 (siblings, not the '346).
• 2024-02-13 — Patent 11,896,346 granted to Omni MedSci, LLC.

---

Timeline diagram

timeline
    title Ownership of US 11896346
    2012 : Provisionals filed Dec 31
    2013 : Islam assigns rights to Omni MedSci Inc
    2018 : Omni sues Apple on family patents
    2023 : Continuation application filed Jan 30
         : Reassignment to Omni MedSci LLC recorded Sep 21
    2024 : Patent 11896346 issues Feb 13
         : Omni sues Samsung Fossil OnePlus Oura

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT (data-limited).
The '346 never moved to a third-party, unrelated licensing vehicle. It has sat with the sole inventor and his own originally-named company throughout. The only entity change is Inc. → LLC (2013 assignee "Omni MedSci, Inc." vs. 2023 assignee "Omni MedSci, LLC"), which is consistent with a corporate-form conversion of the same controlled entity rather than a transfer to an unrelated shell. No evidence of a registered-agent-service address or a single-purpose Delaware/Texas LLC. However, because the 2023-09-21 record lists the assignor as ISLAM, MOHAMMED N. (not Omni MedSci, Inc.), I cannot rule out an interposed LLC structure; that gap requires the recorded document.

2. Known asserter in the chain — UNCLEAR / PARTIAL.
Omni MedSci is a serial patent plaintiff: Apple (2018, 2019, 2020, and a 2026 W.D. Tex. action), Samsung/Fossil/OnePlus/Oura (E.D. Tex. 2024, 2025, 2026), and Whoop. That squarely fits the "repeat high-frequency plaintiff" profile tracked by Unified Patents/RPX. But it does not match the legacy NPE rosters enumerated in the rubric (Acacia, Marathon, IV, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, DGC, Spangenberg entities). I found no source placing Omni MedSci on a published NPE list, so: unclear — do not assert it as a listed NPE without an RPX/Unified citation.

3. Repeat correspondent across the chain — UNCLEAR on the assignment records; PRESENT at prosecution level.
The recurring prosecuted/recording firm for this entire portfolio is Brooks Kushman P.C., 1000 Town Center, 22nd Floor, Southfield, MI 48075, customer name "Brooks, Kushman P.C./Cheetah Omni MedSci," with attorney David S. Bir (Reg. No. on file; signature "/David S. Bir/" appears on family filings; docket e.g. OMN10102PUSA6). This is the tell the task asks for — a single repeat-player firm/attorney behind an anonymously-labeled family. But I could not confirm Brooks Kushman/Bir as the correspondent of record on the 2013 or 2023 assignment instruments; that field is exactly what was not retrievable. And per the rubric's own caution, this firm does substantial operating-company work, so a single appearance is not a finding — here the recurrence is across the whole family, which is why I flag it rather than call it.

4. Cascading transfers — NOT PRESENT.
Only one post-issuance recorded event (2023-09-21) exists; there is no chain of consecutive LLC-to-LLC assignments within 24 months, and no shared-correspondent-address shuffling.

5. Pre-litigation transfer — NOT PRESENT.
The '346 is not verified as asserted in any infringement suit (see the prior litigation section: its only docket appearance found was as Petitioner Exhibit 1088 in Samsung v. Omni MedSci, PGR2025-00064, which was not an infringement assertion). The 2023-09-21 reassignment preceded the earliest sibling-patent Samsung/Fossil complaint (Dec. 20, 2024) by ~15 months — outside the 6-month window — and preceded issuance, not suit.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 proceeding for Omni MedSci, Omni MedSci LLC, or (in the relevant respect) the University of Michigan.

7. Privateering — NOT PRESENT.
Privateering requires an operating company transferring to an NPE to assert on its behalf against competitors. Here the asserting entity is the inventor's own company, which is the original assignee — there is no upstream operating-company sponsor and no SEC-filed divestiture.

8. Defensive aggregator — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. The patent remains with the asserting entity, so it is not neutralized.

---

Verdict

NPE — moderate confidence.

Justification: The '346's title has never left the control of sole inventor Mohammed N. Islam and his wholly-controlled company — the chain runs Islam → Omni MedSci, Inc. (executed 2013-12-17) → Omni MedSci, LLC (recorded 2023-09-21) — and that entity's observable activity is serial assertion against large operating companies (Apple 2018–2026; Samsung, Fossil, OnePlus, Oura 2024–2026; Whoop) with no verified commercial product embodying the claims despite a 2012 priority date and 2024 issuance. That is a genuine NPE signal (rubric #2, "repeat high-frequency plaintiff," and the general "product-less assertion entity" pattern). It is only moderate, not high, because the classic NPE mechanics are absent: there is no shell-LLC cascade, no known-asserter roster listing, no pre-litigation transfer, and no downstream assignee — the patent sits with the original assignee, and Omni MedSci holds itself out as a product company. If you can retrieve the Assignment Center record and it shows an interposed single-purpose LLC or a repeat "licensing" correspondent on both links, upgrade to high; if Omni MedSci is shown to ship an embodying product, the correct call flips to operating-company assertion.

Verify at: https://assignment.uspto.gov/patent/index.html (query "11896346") — and cross-check the family in the USPTO Assignment Center at https://assignmentcenter.uspto.gov/.

I'll research the prior art citations for US 11,896,346 specifically. Let me start with targeted searches.

I found a partial "Referenced Cited" list on Justia. Let me get the complete citation list from Google Patents and USPTO sources.

I need the full "References Cited" list. Let me pull the patent PDF and Justia/FreePatentsOnline entries directly.

I hit the tool step-limit while pulling the authoritative USPTO citation tables, so I could not complete a full register-by-register citation pull. Below I give you (a) exactly what I verified, (b) a clearly-labeled list of the references that appear on the face of the '346 per the Justia mirror of the USPTO front page, and (c) the family-level prior art that is actually the most probative for §102 — with an explicit note on what still needs verification. I have flagged every place where I am inferring rather than citing.

---

1. The patent, identified literally

Flag: Google Patents shows "Family has litigation" with a Darts-IP family link (family=69228103), and a separate docket analysis confirms the '346 is now asserted (Omni MedSci v. Apple, W.D. Tex. 7:26-cv-00226, asserting claims 1, 6 and dependents 2–5, 7–11 against Apple Watch/AirPods/iPhone/iPad "heart rate sensors and skin sensors"). URL: https://ai-lab.exparte.com/case/dct/txwd/7:26-cv-00226/doc/analysis/1. That is a newer fact than the litigation summary in the prior section and should be reconciled there.

---

2. The §102 date framework you must apply first (this governs everything below)

Because the '346 is a continuation filed 2023-01-30 that claims 2012-12-31 priority, prior-art status must be analyzed against the effective filing date of each claim, not the issue date:

• If claims 1/6/15 are entitled to 2012-12-31 → §102 art must predate that (pre-AIA §102(a)/(b)/(e), or AIA §102(a)(1)/(a)(2) if the family crosses the AIA line as Google's "Prior art date 2012-12-31" suggests).
• Critical caveat: the "identify the … user based on water absorption … at ~1090 nm (lower water absorption) vs ~1440 nm (higher water absorption)" limitation is the apparent point of novelty of the '346. If that subject matter is not supported in the 2012 priority application, the claims get a much later effective date (2021 or 2023), which pulls large amounts of post-2012 art into §102 range. I could not verify written-description support within my tool budget; this is the single highest-value check for an invalidity analysis.

Also note: most references that appear on the face of the '346 issued 2013–2018, i.e., after the nominal priority date. They can only be §102 art via an earlier effective filing date (pre-AIA §102(e) / AIA §102(a)(2)). I have issue dates but not filing dates for most of them — see the verification gaps in §5.

---

3. References appearing on the face of the '346 ("Cited U.S. Patents" / §56)

Per the Justia mirror of the granted front page (https://patents.justia.com/patent/11896346), the cited U.S. patent documents are:

Non-patent literature appearing on the face (§56 "Other References"):

• IPR No. 2021-00453, Decision Granting Institution of IPR of U.S. Pat. No. 10,517,484, Apple, Inc. v. Omni MedSci, Inc., pp. 1–50 (Aug. 6, 2021).
• IPR No. IPR2020-00175, Judgment/Final Written Decision of U.S. Pat. No. 10,188,299, pp. 1–71 (Jun. 14, 2021).

Honest limitation: I could retrieve the citations and dates above but could not retrieve the full text of each reference to confirm (i) its filing date or (ii) that it discloses the specific water-absorption identification feature. I therefore will not assert a firm §102 anticipation mapping to these eleven entries; treat the right-hand column as hypotheses to verify, not findings. Two things argue that these face-of-patent references are background/component art (detectors, filters, packaging optics) rather than the anticipatory art for the water-absorption claims: (a) their subject matter is generic photodetector/spectrometer art, and (b) the "1090 nm vs 1440 nm water-absorption identification" concept is the '346's distinctive contribution, which the examiner's own cited art does not appear aimed at.

---

4. The prior art that is actually most probative for the '346 (family-level, from PTAB and invalidity contentions)

The real §102/§103 attack material for the Omni MedSci family — including the sibling patents sharing the '346's specification — comes from Apple's IPR petitions and invalidity contentions. Flag: these references were asserted against sibling patents ('533, '484, '546, '286, '698, '040), not the '346; I have not seen a §102 chart mapping them to the '346's water-absorption limitations.

Core references (from IPR2021-00453 exhibit list, https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1546302](/patent/1546302)/...):

Additional references Apple elected/asserted in the parallel district-court case (https://www.docketalarm.com/cases/PTAB/IPR2020-00029/... and https://www.docketalarm.com/cases/PTAB/IPR2019-00916/...):

• US 8,172,761 (Rulkov) — filed 2011-10-04, issued 2012-05-08.
• US 7,648,463 (Elhag) — issued 2010-01-19.
• US 5,795,300 (Bryars) — issued 1998-08-18.
• US 8,954,135 (Yuen) — filed 2013-06-24 / priority 2012-06-22, issued 2015-02-10.
• US 6,731,967 (Turcott) — issued 2004-05-04.
• US 8,755,871 (Weng) — filed 2011-11-30, issued 2014-06-17.
• US 8,315,682 (Such) — filed 2005-12-05, issued 2012-11-20.
• US 6,708,048 (Chance) — NIR tissue optical examination; note Chance is the strongest candidate for claim 6/15 (NIR/SWIR tissue spectroscopy, selective wavelengths).
• Systems: Masimo Radical-7, Masimo iSpO2, Nellcor, Nonin pulse oximeters; Fitbit; Asada.
• Apple's own later art pulled in via the '346 claim chart: US 2022/0225060 A1 (Allee et al., Apple, "Electronic Devices With Skin Sensors") and US 11,678,805 B2 — URL: https://storage.courtlistener.com/recap/gov.uscourts.txwd.1172913743/gov.uscourts.txwd.1172913743.1.12.pdf.

I did not find (and will not invent) a reference that squarely discloses the '346's signature pair "modulated LED near 1090 nm (low water absorption) + modulated LED near 1440 nm (high water absorption), compare the two reflected signals to identify the object." The closest thematic art I encountered in searching is WO 2017/187088 A1 (PCT/FR2017/050992) — water-concentration measurement in a diffusing material using sources at water-absorption peaks (~1450/1500/1550 nm) vs non-absorption (~1300 nm) and source/detector pairs, with a patch/holder — but (a) I did not verify it is cited against the '346, and (b) it postdates the 2012-12-31 priority date, so it is §102 art only if the water-absorption claims lose the 2012 priority. It is nonetheless a high-value lead if the priority challenge succeeds.

---

5. Verification gaps I could not close (do not treat this section as findings)

1. Full USPTO "References Cited" pull — I reached the tool step-limit before querying USPTO Patent Public Search / PatentCenter / the patent PDF directly. The §3 list is from the Justia mirror; it should be re-confirmed against the granted PDF (patentimages) and the file-wrapper IDS to catch any entries the mirror truncated (the Justia snippet contained ellipses, so it may be incomplete).
2. Filing dates of the §3 references — needed to determine which of them are even eligible §102 art (pre-AIA §102(e) / AIA §102(a)(2)).
3. Substance of each §3 reference — needed before any specific claim can be said to be "anticipated." I have deliberately not fabricated claim mappings for these.
4. Effective-date / written-description support for the water-absorption limitations in the 2012 provisional (61/747,472) and the intervening applications — decisive for how much of the post-2012 art becomes available.
5. Whether the '346 is in any IPR/PGR on its own claims — per the prior litigation section, the only confirmed PTAB appearance is as Exhibit 1088 in PGR2025-00064 (challenging the '790), i.e. not a challenge to the '346.

---

Bottom line

• I confirmed the bibliographic identity and the family/priority structure of US 11,896,346 and confirmed that it is now asserted against Apple in W.D. Tex. 7:26-cv-00226 (claims 1, 6, 2–5, 7–11) — a newer fact than the litigation summary, which should be updated.
• I identified the eleven U.S. patents cited on the face of the '346 and its two cited PTAB decisions (Section 3), with dates — but could not verify their substance or filing dates, so I have not asserted firm §102 anticipation for them.
• The honestly most probative §102/§103 art is the family-level prior art used in Apple's IPRs and invalidity contentions (Valencell-093, Hanna, Mannheimer, Carlson, Park, Lisogurski, Chance, Rulkov, Elhag, Bryars, Yuen, Turcott, Weng, Such, plus the Masimo/Nellcor/Nonin/Fitbit systems) — Section 4 — but that art has not been mapped to the '346's water-absorption claims.
• The single most consequential open question is priority: if the "identify based on water absorption (~1090 nm vs ~1440 nm)" claims are not supported by the 2012 priority application, the effective date moves to ~2021–2023 and substantially more art (including the WO 2017/187088-type water-content art) becomes available.

If you want, I can, on a fresh run: (1) pull the exact §56 table and IDS from the USPTO file wrapper for 18/103,408, and (2) build a claim-chart mapping of the Section 4 references against the '346's independent claims 1, 6 and 15.

I'll ground this analysis in the actual prior-art record for the '346 patent. First, let me try to retrieve the cited references.

I found several concrete prior-art leads, including a Japanese publication expressly discussing 1200/1439 nm water-absorption bands and a WO publication on determining tissue water content. Let me pin down the cited references and their dates.

Source and verification note (read first)

The task says to use "the results from the Prior Art section of this page." The Google Patents page text I was given contains no "References Cited"/"Citations"/"Cited By" list — the fetch was truncated after the figure list, and all that survives of the prior-art metadata is the auto-generated field:

Prior art keywords: light; wavelength; output; fiber; optical

and the "Definitions" concept list (non-invasive glucose monitoring, near-infrared spectroscopy, spectral fingerprinting, dental caries, SWIR, eye-safe window, etc.). I therefore could not satisfy the instruction literally. What I did instead, and what I label below:

⚠️ Contradiction flag (per my cross-reference instructions). My two previously generated sections disagree: the Patent summary says the '346 is asserted/charted in Omni MedSci v. Apple, W.D. Tex. 7:26‑cv‑00226; the Litigation summary says the '346 is only a "family relative" and that the '475 and '787 are the asserted patents. The new search evidence sides with the Patent summary: the CourtListener PDF for the W.D. Tex. docket contains a document captioned "U.S. Patent No. 11,896,346 (the '346 Patent)" with a two-column chart headed "346 Patent Claim / Apple Accused System" and element labels 1[1], 1[n], 4[b] — i.e., a claim‑by‑claim infringement chart for the '346. I cannot resolve which is correct; treat the '346 as possibly asserted in the 2026 Apple case and verify on the live docket. This matters for the §103 analysis only to the extent it tells us which claims the parties are actually fighting (1 and 4 at minimum).

---

§103 Obviousness Analysis — US 11,896,346 B2

1. Threshold: the effective filing date controls the entire analysis

This is the single most consequential issue, and it is a live one in this family.

• The '346 issued from application 18/103,408, filed 30 January 2023, as a continuation claiming priority back to 31 December 2012 (prov. 61/747,485 / 61/747,477) via the 17 Dec 2013 non‑provisionals and PCTs.
• It is therefore an AIA patent. Under §102/§103 the critical date is the effective filing date of the claimed subject matter.
• Scenario 1 — Dec 31 2012 priority holds. The prior-art universe closes in on 2008–2012 art. The references in §2(a)–(c) are still squarely §102(b)/§103 art. Apple's Allee publication (July 2022) and WO 2017/187088 / US 2019/0137385 are not available.
• Scenario 2 — priority fails for these claims under §112 (no written-description support in the 2012 provisional/PCT for the specific "identify … based on water content" dual-wavelength + threshold architecture). Then the effective date is 30 January 2023, and every 2022/2019/2017 publication below becomes §102(a)(1) art. I previously flagged that the family has already drawn an AIA/effective-filing challenge (a "Director Squires"-type issue in the '790/PGR track), and the W.D. Tex. chart is built almost entirely on July 2022 Apple publications — a strong tell that the accused-infringer narrative assumes a 2023 effective date for the water-content claims.

Analyst position: the claims are obvious in both scenarios, but the combination differs. I give a combination for each.

PHOSITA level: a B.S. in optics/biomedical engineering (or equivalent) with 2–5 years' experience designing NIR/SWIR reflectance oximetry or hydration instrumentation, including LED/laser‑diode selection, dielectric/bandpass filtering, modulated (lock‑in) detection, and chemometric signal processing. This is the level Apple's experts (e.g., MacFarlane, in the sibling IPRs) have applied in this family.

---

2. The prior art

(a) WO 2008/112136 A1 — "System and method for tissue hydration estimation" (PCT/US2008/003015; published 18 Sept 2008)

This is the closest single reference and it reads on most of the architecture. Verified claim text:

• Claim 1: "transmitting electromagnetic radiation at tissue; detecting the absorption spectrum of the tissue using a spectrum analyzer located in a sensor; … processing the signal to determine an amount of water content in the tissue."
• Claim 2: "transmitting a plurality of discrete wavelengths within the near-infrared (NIR) spectrum."
• Claim 4: "processing the signal comprises analyzing the distribution of spectral power to determine a ratio of water to other constituents."
• Claim 8/11: "filters to control the bandwidth of electromagnetic radiation that impinges on a detector array"; a MEMS Fabry‑Perot filter.
• Claim 14/15: "a plurality of narrow band light sources … light emitting diodes (LEDs) operating in the NIR band."
• Auto-extracted keywords: electromagnetic radiation; tissue; spectrum; water; spectrum analyzer.

That is: NIR LEDs + spectrally selective filtering + a detector + processing that computes a water ratio and displays it. Everything in claim 1 except the specific 1090/1440 nm pair, the modulation/synchronization, the wearable housing, and the word "identify."

(b) JP 2009‑538672 A / Philips PCT family — "Apparatus for depth‑resolved measurement of tissue attributes" (priority 30 May 2006; JP publication 12 Nov 2009)

This is the reference that most directly prefigures the 1090/1440 nm pair as a water-absorption contrast measurement. From the document (machine-translated excerpts retrieved):

• The light source "preferably has wavelengths of … 758 nm, 914 nm, 1200 nm, 1439 nm, 1932 nm … and/or between about 1600–1900 nm."
• Figs. 9–12 plot diffuse-reflectance spectra "around the water spectral bands, i.e., around 1200 nm and 1439 nm," where "clear differences in intensity are visible. Thus, from these spectra, the water content and changes in water content in the object — in this example the tissue — can be determined."
• The difference spectrum at ablation shows the reflectance increase "indicates a decrease in water absorption at these spectral bands, and hence a decrease in water content."
• "The NIR spectra shown in Figs. 9–12 and 15 were acquired using extended InGaAs photodetectors."

So a single reference teaches: (i) illuminating tissue at a lower‑water‑absorption NIR wavelength (~1200 nm) and a higher‑water‑absorption NIR wavelength (~1439 nm); (ii) detecting the reflected light with InGaAs; and (iii) comparing the two returns to determine water content. Claim 1's "second wavelength with higher water absorption near 1440 nanometers" is a near-literal hit (1439 nm). The 1200 → 1090 nm shift is the only meaningful difference, and it is a §103 design choice, not a §102 gap (see §3(c)).

(c) US 8,180,419 B2 — "Tissue hydration estimation by spectral absorption bandwidth measurement" (Nellcor Puritan Bennett LLC)

Verified specification text: the emitter "may be at least one device capable of emitting electromagnetic radiation in the 980 to 2350 nm spectral region … the emitter may be one or more LEDs operating at wavelengths indicated in Table 1"; the detector "may be an InGaAs detector"; a processor correlates the detected light to a hydration index; the device is explicitly handheld/packable ("the system 50 may be configured as a single unit that can be carried with a user, nurse or doctor") with a housing (opaque underside with emitter and detector apertures) and a display.

This reference supplies: LEDs in the water‑band region, InGaAs detection, tissue water quantification, a handheld housed form factor, and thresholding/notification (displaying a hydration index or dehydration condition).

(d) WO 2017/187088 A1 (+ apparent US counterpart US 2019/0137385 A1) — "Method and apparatus for measuring the water concentration in a light‑diffusing material"

Verified content: a source "centred around 1450" nm and a source at "a zero‑water‑absorption wavelength"; it expressly criticises the prior art because "these techniques lay on the assumption that absorption by water of radiation emitted at the wavelength(s) distant from the absorption bands is negligible" — i.e., it confirms the two‑wavelength water‑contrast ratio was already the state of the art. It also (i) is a wearable/portable implementation, and (ii) lists intended uses including "characterising the hydration of the skin … the water concentration in a food matrix … in a medicinal product … in a hydrocarbon … in a construction material."

That last list matters enormously: it is the **"user" and "object" **framing of claims 1 and 6 in a single reference, and it maps directly onto the '346's own field (skin, food/agriculture, illicit/counterfeit drugs, plastics, minerals/gas). If the effective date is 2023, this is devastating art.

(e) US 2022/0225060 A1 — Allee et al., "Electronic Devices With Skin Sensors" (Apple; app. 17/149,552, filed 14 Jan 2021, published 14 July 2022)

Verified abstract: "A wearable electronic device such as an earbud, wristwatch, or other device may be provided with a skin sensor … first and second light-emitting devices such as infrared devices that emit light at respective first and second infrared light wavelengths. Reflected light is monitored by a photodetector. Control circuitry can initiate or pause audio playback or take other actions in response to determining from the reflected light measurements that skin is present. … The sensor may have a thin-film interference filter or other optical structure that overlaps the first and second light-emitting devices to narrow the angular spread … This reduces tilt sensitivity."

Formatting note (per the "do not auto-correct" rule): the same court exhibit renders this reference under two numbers — "US 2022/0225060 A1" and "US 2022/0225006 A1" — and its element column cites the underlying PDF as US20220225006A1.pdf. Both strings appear in the record; I do not reconcile them.

Evidentiary caveat: in the complaint this publication is used as proof of the accused Apple device, not as an invalidity citation. But if the effective filing date is 30 Jan 2023, it is a printed publication under §102(a)(1) and can be used as prior art with no such caveat.

(f) Background/practitioner art confirming the POSITA's toolkit

• JP 2009‑515664 A, "System and method for assessing one or more body fluid indices" — "at least two different wavelengths … the absorption of light at the at least two different wavelengths is mainly due to water in vascular blood and extravascular tissue, and the ratio of the at least two measurements … provides a measure of the difference in moisture content"; probe housing + LEDs + fiber delivery. This is the ratio‑of‑two‑water‑band‑returns technique, in a housed probe, in 2009.
• The search report published inside WO 2008/112136 cites WO 02/26152 (Instrumentation Metrics), US 2004/0230106 A1 (Schmitt et al.), WO 2005/038418 (Aysun Tech / Atia), EP 1 184 663 (Spectron), JP 2004‑081427, WO 2006/020408 (Glucolight / Schurman & Shakespeare) — an examiner‑vetted snapshot of the NIR analyte‑measurement toolkit as of 2008. (These are the WO'136 ISR hits; I did not obtain the '346's own IDS list — see caveat §6.)
• The sibling Apple IPR (IPR2021‑00453 on the '484, petition text retrieved) asserts that LED + lens + receiver + "standard signal processing" devices for physiological parameters are "an entirely predictable combination of well‑known components," relying on Lisogurski, Carlson, Tran, Isaacson and Park — the same evidence base a POSITA in this family would use to show modulated/synchronized detection was routine.
• Wang et al., "Multichannel Reflective PPG Earpiece Sensor with Passive Motion Cancellation," IEEE Trans. Biomed. Circuits & Syst., Dec. 2007 (listed in this family's own IDS) — a housed ear‑worn reflectance sensor with synchronized/motion‑cancelled detection.

---

3. Element-by-element analysis

Claim 1 ("sensor system … identify [a] user based on water absorption")

Claim 1 is the weakest claim in the set. Allee alone supplies the wearable two‑IR‑LED + filter + photodetector + determination‑against‑criterion architecture; JP 2009‑538672 alone supplies the 1200/1439 nm water‑contrast measurement; WO'136 supplies NIR LED + filtered detection + water‑ratio processing. The only residual novelty is the label — calling the output an "identification" of a "user" rather than a "presence"/"hydration" determination.

Claim 6 ("optical system … identify [an] object," + threshold comparison)

Same map, plus:

• "modulated light‑emitting diode" recitations → trivial (WO'136 LEDs; Allee IR LEDs);
• "compare the output value to a threshold" → Allee (control circuitry "initiate or pause audio playback … in response to determining from the reflected light measurements that skin is present" necessarily applies a decision criterion); '419 (hydration index thresholds / "over‑hydration … dehydration" determinations); WO 2019/0137385 (calibration/model comparison). Comparing a computed value to a threshold is the sine qua non of any binary optical determination and is at the "mere design choice" end of the spectrum.

Where the '346's own specification concedes the point

The specification supplies §103 ammunition against its own claims:

• It states that "spectroscopy and pattern matching [is] often called spectral fingerprinting" and treats it as known;
• It concedes that the eye-safe window (>1400 nm) and the water transmission windows (1500–1850, 2050–2500 nm) are known;
• It concedes that "SWIR light may be generated by light sources such as lamps, [LEDs], one or more laser diodes, super‑luminescent laser diodes, and fiber‑based super‑continuum sources" — i.e., the diode selection is conventional;
• It labels the wavelength teaching "exemplary … other ranges can be used," which undercuts any argument that 1090/1440 nm is a critical, non‑obvious selection.

---

4. Specific §103 combinations and motivations

Combination A — Scenario 1 (2012 priority holds): WO 2008/112136 + JP 2009‑538672 + '419

Motivation to combine.

1. Same field, same problem. All three address non‑invasive optical determination of tissue water/hydration by NIR reflectance. A POSITA working on wearable body sensors in 2012 would routinely consult this literature.
2. Complementary disclosures, no functional interference. WO'136 teaches the architecture (NIR LEDs → spectrally filtered detection → processor computes a water ratio), but its claim 2/14–15 covers "a plurality of discrete wavelengths" without naming them. JP 2009‑538672 supplies precisely those wavelengths and the analytic justification — 1200 nm (low water absorption) vs. 1439 nm (high water absorption) — and even specifies InGaAs detection, matching WO'136's detector‑array concept and '419's express InGaAs teaching.
3. Predictable result (KSR rationale (1)). Combining a known water‑ratio computation with known water‑band wavelengths yields nothing more than the expected improvement in water‑contrast specificity. There is no teaching away: the '346's own Fig. 3B shows hemoglobin/oxygenated‑hemoglobin features below ~1100 nm complicate measurement, which is a reason to prefer ~1200 nm (JP'672) — i.e., the prior art's choice is at least as good as the claim's.
4. '419 supplies the handheld, housed, display‑and‑threshold implementation, which supplies the "housing … coupled to an electrical circuit and a processor" and the claim‑6 threshold step, and confirms InGaAs for the detector dependent claim.

Result: claims 1 and 6, and the InGaAs, housing, filter, threshold, and WBC/water‑ratio dependents are obvious.

Combination B — Scenario 2 (effective date 2023): Allee (US 2022/0225060) in view of WO 2017/187088 / US 2019/0137385 (and optionally '419 or WO'136)

Motivation to combine.

1. Both are directed to wearable/portable optical sensing of a target's water content — Allee to "detect the presence of human skin" in an earbud/watch; WO'187088 to "characterising the hydration of the skin" and the water concentration in "a food matrix … a medicinal product … a hydrocarbon." Same field, overlapping problem (reject false positives; classify the target).
2. Allee supplies every structural element of claims 1 and 6: plurality of modulated IR LEDs, thin‑film interference filter overlapping the emitters, photodetector, wearable housing, control circuitry, and a determination against a criterion with an action taken.
3. WO'187088 / US 2019/0137385 supply the missing analytic core — emit at a water‑absorption wavelength (~1450 nm) and at a zero‑water‑absorption wavelength, receive reflected light, and compute the water concentration, with the express caveat that the prior art already assumed the distant wavelength's water absorption is negligible. That is the claimed "compare the first part and the second part to generate an output value."
4. KSR rationale (4)/(5). Improving a known skin‑presence sensor by using water absorption as the discriminating physical mechanism is "applying a known technique to a known device ready for improvement," and the water‑band wavelengths are among a finite number of identified, predictable solutions (the patent itself names only ~1090 and ~1440 nm). WO'187088's λ ≈ 1450 nm is within rounding of the claimed "near 1440."
5. Apple's own chart is evidence of predictability. The W.D. Tex. exhibit maps the '346's claim 1, claim 4, and the threshold element element‑by‑element onto Apple's shipped AirPods skin sensor, quoting Allee for each: two IR LEDs, thin‑film interference filter, photodetector, and a compare‑to‑threshold determination. A claim that reads on a 50‑cent, mass‑produced consumer part described in a July 2022 publication is, on its face, the product of ordinary skill rather than invention.

Result: claims 1 and 6 and essentially all dependents are obvious; the "identify based on water absorption" limitation is met by WO'187088's water‑concentration computation performed in a device whose explicit purpose is to distinguish/characterise the target.

Combination C — JP 2009‑538672 alone, or + a secondary teaching of "identification"

JP'672 already discloses (i) illumination at wavelengths including 1200 nm and 1439 nm, (ii) detection with extended InGaAs, and (iii) computing water content and changes in water content from the intensity contrast at those bands. If the claims are construed so that "identify … based on water absorption" requires no more than generating a water‑content value used to characterise the target, claim 1 falls to this reference alone plus the ordinary creativity of adding a modulated source and a threshold. The remaining gap (the recitation of "identifying a user/object" and the 1090 nm value) is met by:

• JP'664 (ratio of two water‑band returns gives a moisture value in a housed probe);
• Allee/‘419 (threshold‑based determination triggering an action).

Range/optimisation argument: under In re Peterson / In re Applied Materials, selecting ~1090 nm for the "low water absorption" channel when the prior art teaches the function at 1200 nm, and ~1440 nm when the prior art teaches 1439 nm, is at most routine optimisation of a result‑effective variable with a predictable effect (less water absorption at the shorter wavelength). The specification offers no data showing a criticality or unexpected result for 1090 nm versus 1200 nm — it merely calls its ranges "exemplary."

Combination D — "Teeth," "packaging/enamel/dentine" dependents

If the effective date is 2023, the family's own 2013–2014 publications — e.g., US 9,164,032 and the PCT/US2013/075736 dental‑caries work, and US 2014/0188092 — are available as prior art in the §102(a)(1) sense to the extent they are not excepted by common ownership under §102(b)(2)(C) (which applies only to §102(a)(2) art). Independently, the "encapsulating layer = enamel, dentine, plastics or packaging material" dependents are squarely met by the through‑package NIR spectroscopy literature the '346's own background recites (drug screening "through the packaging non‑destructively"), and the SWIR dental literature (trans‑illumination of teeth, QLF, 1/λ³ scattering). A POSITA would combine these because the claimed subject matter is simply the application of a known water‑contrast measurement to an article with known SWIR transparency.

---

5. Dependent claims — brief mapping

Note that "modulate," "modulating," "one or more filters," and "to identify an object" were all litigated as claim‑construction terms in the E.D. Tex. Samsung/Omni case (per the Nov. 2025 joint claim‑construction service email), and "identify an object" was addressed by the Federal Circuit in Apple v. Omni MedSci, No. 23‑1034 (June 21, 2024). The §103 outcome for claims 1 and 6 is materially sensitive to how "identify … based on water absorption" is construed — if it requires biometric identification of a specific individual, WO'136/'419/'672 (hydration meters) do not disclose it and the analysis shifts entirely onto Scenario 2 (Allee + WO'187088). If it means "characterise what the target is / that the target is present," the art in Combination A/C is dispositive.

---

6. Rebuttal considerations (objective indicia and teaching away)

• Secondary considerations. I have found no evidence of unexpected results, long‑felt unmet need, industry praise, licensing, or failure of others tied by nexus to the 1090/1440 nm selection. Any such showing would have to attach to the specific wavelength pair rather than to the general commercial success of SWIR sensing. The patent's own "exemplary … other ranges can be used" language makes nexus difficult.
• Teaching away. The specification says the "advantageous window for measuring glucose and other blood constituents may be in the SWIR between 1500 and 1850 nm and 2050 to 2500 nm" and that below ~1100 nm hemoglobin features "may make it more difficult to measure blood constituents." A patentee could argue this teaches away from a 1090 nm channel. Assessment: weak. (i) The statement is about glucose measurement sensitivity, not about a water‑content discriminator; (ii) it is a statement of difficulty, not of inoperability (In re Fulton; "criticism of a reference does not teach away if it does not indicate the invention should work differently"); (iii) the "eye safe" discussion (>1400 nm) actually supports a 1440 nm channel.
• Claim‑construction risk. As above — the narrowest plausible reading of "identify" is the patentee's best defence in Scenario 1, but it is self‑defeating in Scenario 2, where Allee‑plus‑WO'187088 shows the same result achieved in a consumer wearable.
• The strongest §103 proof point remains the priority gate. If the "water content identification" limitations are new matter relative to the 31 Dec 2012 priority documents, the claim's effective date is 30 Jan 2023 and the art expands to include Apple's July 2022 Allee publication — which the patentee's own litigation opponent has already charted as reading on claims 1 and 4.

---

7. Bottom line

1. Claim 1 (user) and claim 6 (object) are, on the current record, strong §103 targets — claim 1 more so than claim 6, because claim 6's "compare the output value to a threshold" is squarely and expressly met by Allee and '419.
2. The single most on‑point reference is JP 2009‑538672 A (Philips family, priority 30 May 2006), which expressly teaches measuring water content from reflectance contrast at 1200 nm and 1439 nm using InGaAs detectors. It largely anticipates the analytic core of the claims and leaves only the "identify" label and the 1090 nm value.
3. WO 2008/112136 A1 supplies the claim architecture (NIR LEDs, spectrally selective filtering, detector array, processing to a water ratio), and US 8,180,419 B2 supplies the housed handheld form factor, LEDs across 980–2350 nm, and express InGaAs detection.
4. If the effective filing date is 30 January 2023, US 2022/0225060 A1 (Allee, "Electronic Devices With Skin Sensors") + WO 2017/187088 / US 2019/0137385 form a nearly complete, well‑motivated combination covering both the structure (wearable, two IR LEDs, thin‑film interference filter, photodetector, threshold logic) and the water‑contrast analysis.
5. Anticipated rebuttals (narrow construction of "identify"; the specification's preference for 1500–1850 nm) are weak and/or self‑limiting, and the specification's own admissions (spectral fingerprinting known; wavelength ranges "exemplary") cut against non‑obviousness.

Verification gaps I could not close

• The '346's actual "References Cited" / IDS list was not in the fetched page text and I could not retrieve it. Combinations A–C are therefore built from references I located independently (and from the WO'136 ISR), not from the '346's own examiner citation list. If the '346's IDS contains, e.g., the Philips JP'672 family or the Nellcor '419 patent, the obviousness case strengthens considerably; if it does not, expect a §103(c)/KSR argument to hinge entirely on the motivation analysis.
• Exact publication dates for JP 2009‑538672 A, US 8,180,419 B2, JP 2009‑515664 A and WO 2017/187088 A1 are stated from secondary records and should be confirmed against the front pages.
• Whether the '346 is actually asserted in Omni MedSci v. Apple, 7:26‑cv‑00226 (contradiction flagged in §0).
• Applicability of US 2019/0137385 A1 as the US member of the WO'187088 family is inferential (the paragraph text matches closely); confirm via the family record.
• The Fed. Cir.'s holding on "identify an object" in No. 23‑1034 would materially sharpen the claim‑scope discussion; I did not retrieve the opinion text here.