US 12588820

I'll research this specific patent number across the sources available to me.

Let me pull the claim language and litigation details more precisely.

US Patent 12,588,820 — Research Summary

Caveat up front: I could not retrieve the issued claim set verbatim for US 12,588,820. The authoritative text I have for this patent stops in the Detailed Description (mid-Section 1 on non-invasive glucose). Search results I found consistently referenced the sibling/specification-sharing patents (the ’475, ’790, ’484, ’304, ’455, ’533) rather than the ’820 itself. Claim characterizations below come from the patent's own SUMMARY section and should be treated as high-confidence but not verbatim claim text. I also found no Federal Circuit 2026 docket naming US 12,588,820 specifically — see the CAFC note at the end.

---

Bibliographic Data (per the authoritative patent text)

Continuation chain (from the patent's own CROSS-REFERENCE section): This application is a continuation of 18/891,970 (filed Sep. 20, 2024), which continues back through 18/438,144 (now US 12,193,790), 18/103,408 (US 11,896,346), 17/514,778 (US 11,564,577), 17/078,771 (US 11,160,455), 16/722,188 (US 10,820,807), 16/506,885 (US 10,517,484), and further parents rooted in the Dec. 17, 2013 PCT/US2013/075700 filing and Dec. 31, 2012 provisionals.

Classifications (representative): A61B5/00, A61B5/0002, A61B5/14532, A61B5/1455, A61B5/681/682, A61B5/7257 (Fourier transforms), G01J3/0218 (optical fibers in spectrometers), G16H40/63, G16H50/20, A61C19/04, H01S3/067 (fiber lasers).

---

Abstract (verbatim)

"A wearable device to measure a user's physiological parameters comprising one or more biosensors, as well as a light source comprising light emitting diodes, lenses for directing light towards tissue of the user comprising blood vessels, and a detection system receiving reflected tissue light. The physiological parameters, for example hypertension, are measured with a differential measurement. For example, the physiological parameters may be associated with pulse rate and blood flow. The output signal is associated with the physiological parameters, and artificial intelligence may be used in making decisions regarding the output signal. Signal-to-noise ratio of the output signal may be improved by synchronizing the detection system to the light source, increasing light intensity, and detecting a change. The wearable device is configured to determine that is being worn by the user and may be configured to communicate with a smartphone or tablet."

---

Independent Claims — Plain-Language Overview

Based on the SUMMARY section, the patent appears to have three independent claims (numbers not confirmed in my source material — treat "Claim 1 / Claim 2 / Claim 3" designations as illustrative):

1. Apparatus adapted to be worn by a user

• One or more biosensors placed on the user, measuring physiological parameters via a differential measurement — specifically pulse rate monitoring and blood flow measurement.
• A light source of multiple LEDs producing output optical light at one or more wavelengths.
• One or more lenses receiving that light and delivering it to tissue comprising skin.
• A detection system receiving reflected light and generating an output signal with a signal-to-noise ratio (SNR). It is (a) synchronized to the light source, (b) has multiple spatially separated detectors, and (c) has at least one analog-to-digital converter coupled to a detector.
• SNR is improved two ways: by increasing the LED light intensity from an initial level, and by comparing a "first signal" (LEDs off) with a "second signal" (at least one LED on) — i.e., dark/ambient subtraction.
• The apparatus is configured to determine, from the output signal, that it is being worn by the user.

2. Wearable device worn by a user

• A light source that can be on or off; when on, it generates output light.
• A lens directing at least part of that light toward the user's bodily tissue; and a detector.
• A processor configured to: (1) with the light on and the detector receiving reflected light, generate a first output signal with a first SNR; (2) with the light off and the detector receiving ambient light, generate a second output signal with a second SNR; (3) generate a third output signal from portions of the first and second output signals, having a third SNR greater than both the first and second SNRs, associated with a physiological parameter; and (4) determine from that third output signal that the device is being worn by the user.
• Note: as printed in the SUMMARY, this claim's transitional recitation is grammatically garbled ("...towards a bodily tissue of the user. a detector; and The processor is configured to..."). This may be an artifact of the text as fetched; the operative scope is as summarized above.

3. System for a device worn by a user

• A multi-LED light source producing output optical light; lens(es) directing light to bodily tissue.
• A detection system synchronized to the light source, receiving reflected light and generating an output signal with an SNR.
• SNR increased by raising LED intensity from an initial intensity.
• The detection system has spatially separated detectors and at least one ADC coupled to at least one detector.
• The output signal is associated with one or more physiological parameters and the system measures them when worn.
• The physiological parameters change in response to stretching of the user's hand, or movement of a finger or thumb — an explicit tie to wrist/hand-worn sensing.
• The device is configured to determine, at least partly from the output signal, that it is being worn by the user.

Common inventive thread across all three: LED-based optical sensing at skin, synchronized detection with spatially separated detectors + ADC, and a two-measurement (on/off) differential/dark-subtraction scheme to boost SNR — plus a wearer-detection function derived from the physiological output signal.

---

Litigation and Post-Grant Status

District court (per Google Patents litigation flag and docket search):

• Omni MedSci, Inc. v. [[Samsung Electronics Co.](/litigations/by-defendant/Samsung%20Electronics%20Co.) Ltd.](/litigations/by-plaintiff/Samsung%20Electronics%20Co.%20Ltd.) et al., No. 2:26-cv-00356 (E.D. Tex.) — the case flagged on the patent page; complaint filed ca. Jan. 2026 in Judge Rodney Gilstrap's court. Accused product families span smartwatches, rings, phones, and tablets.
• Related Omni MedSci filings in the same family/campaign include v. Fossil Group (2:26-cv-00470, E.D. Tex.), v. Apple Inc. (7:26-cv-00226, W.D. Tex.), and v. Whoop, Inc. (1:25-cv-00140, D. Del.).
• Note: Omni MedSci's campaign pleadings I found assert the sibling ’475 (12,268,475), ’790 (12,193,790), ’484, ’304, ’455, ’533, and ’868 patents. I did not find a pleading that expressly asserts the ’820 itself — the ’820 is a very recent grant (March 31, 2026) and may be too new to have appeared in an assertion of record in my sources.

PTAB: Samsung (with Fossil, Oura, OnePlus) filed a coordinated wave of IPR/EPR/PGR petitions (e.g., IPR2025-01249 through 01253) against the family. Apple previously challenged the ’533 (IPR2019-00916, all challenged claims held unpatentable, affirmed by the Federal Circuit at 2022 WL 2062168) and the ’484 (IPR2021-00453). Law360 items dated May 2026 report the PTAB declining to run an IPR on an Omni MedSci wearable-device patent because a separate post-grant review would proceed, and a Director decision on the AIA effective-filing-date cutoff for the Omni MedSci patent.

Federal Circuit / CAFC 2026: I found no CAFC docket in 2026 listing US 12,588,820 as the patent on appeal. The relevant Federal Circuit activity is on related patents:

• Apple Inc., et al. v. Omni MedSci, Inc., No. 2023-1034 — "AFFIRMED-IN-PART, VACATED-IN-PART AND REMANDED" (Fed. Cir. June 21, 2024), addressing the ’484 patent's "identify/detect an object" limitations.
• Omni MedSci appeal of the ’484 remand decision, No. 2025-1646 (Fed. Cir.) — reported as pending.

Uncertainty: Because the ’820 issued only on March 31, 2026, any appeal or PTAB proceeding specifically about the ’820 would be at a very early stage. I cannot confirm from my sources that the ’820 is yet the subject of any CAFC appeal or PTAB trial. If you need that confirmed to a legal certainty, the authoritative check is the USPTO PatentCenter/PTAB portal and PACER for E.D. Tex. 2:26-cv-00356 — the patent number itself should be read literally as 12,588,820, and I did not substitute any similar number.

I'll research litigation specifically involving US 12,588,820 using multiple sources.

Let me search more specifically for the case number flagged on the patent and Omni MedSci's recent filings.

I found the key case. Let me verify whether other cases in Omni MedSci's campaign also assert the '820 specifically.

Litigation Involving U.S. Patent No. 12,588,820

Note on the record date: Your task header says April 26, 2026, but the operating date for this analysis is September 28, 2026, and several sources I retrieved are dated May–August 2026. I report everything I found, including post-April events, and flag the discrepancy.

Important correction to my earlier section: In my prior research summary I wrote that I had not found a pleading expressly asserting the '820. That is now superseded — the '820 is asserted, in the Samsung action below. My earlier note that the Google Patents litigation flag pointed to a complaint "filed ca. Jan. 2026" is also wrong on the date: PACER shows May 1, 2026.

---

1. Only confirmed case asserting US 12,588,820

The '820 in this complaint

• The complaint asserts eight patents: 10,126,283; 11,241,156; 11,596,311; 11,678,805; 11,992,291; 12,426,788; 12,588,820; and 12,599,305. (Confirmed in both the complaint analysis and a Korean trade-press report of the filing.)
• For the '820 specifically, Omni asserts at least claims 9 and 10 (Compl. ¶ 232). The '820 is described in the complaint as directed to "a measurement system adapted to be worn by a user" with laser diodes, a photodiode array, first and second cameras, and combined time-of-flight/image data (Compl. ¶¶ 117, 124–126).
• Accused instrumentalities tied to the '820: Samsung's Galaxy XR headset, on the theory that it performs direct time-of-flight (dToF) sensing (Compl. ¶¶ 163, 232). The broader complaint also accuses the laser autofocus (laser AF) feature of Galaxy Note 20 Ultra, S21 Ultra, S22 Ultra, S23 Ultra, S24 Ultra, and S26 Ultra smartphones under other asserted patents.
• Relief sought: declaratory judgment of infringement, damages, permanent injunction, and attorney's fees.

Contradiction flagged: My earlier section characterized the case flagged on the patent page as an Omni MedSci wearable-health campaign (smartwatches, rings, phones, tablets, accused products spanning four brands). The actual 2:26-cv-00356 complaint is a laser-sensor / dToF-focused action against Samsung only, and it asserts a different patent set (the '283, '156, '311, '805, '291, '788, '820, '305 patents) than the wearable-health family asserted in the other Omni suits. The earlier "smartwatch campaign" description more accurately matches 2:24-cv-01070 and the related 2025–2026 filings — not this one.

---

2. Related Omni MedSci litigation that does not (as far as I can confirm) assert the '820

These are part of the same owner's portfolio campaign but assert different patents. I list them only to show the '820 was searched for and not found in them:

I was unable to confirm the complete asserted-patent list for the Apple (7:26-cv-00226) and OnePlus (2:26-cv-00472) complaints, so I cannot rule out the '820 in those. The one case I positively confirmed asserts the '820 is 2:26-cv-00356.

---

3. PTAB and appellate activity specific to the '820

• No PTAB proceeding (IPR/PGR) naming US 12,588,820 was found. The Samsung/Fossil/Oura/OnePlus coordinated petitions (IPR2025-01249 through 01253, PGR2025-00063/00064) target the '533, '484, '304, '455, '790, '475, and '868 patents — not the '820.
• No Federal Circuit appeal listing US 12,588,820 was found. The pending CAFC activity concerns related patents: Apple v. Omni MedSci, No. 2023-1034 (Fed. Cir. June 21, 2024, aff'd-in-part/vacated-in-part/remanded, '484 IPR) and Omni MedSci v. Apple, No. 2025-1646 (pending, '484 remand appeal).
• The '820 issued March 31, 2026 — less than five months before this analysis — which is consistent with the absence of any separate validity challenge.

---

4. Bottom line

• Exactly one known litigation asserts US 12,588,820: Omni MedSci, Inc. v. Samsung Electronics Co., Ltd. et al., No. 2:26-cv-00356 (E.D. Tex.), filed May 1, 2026, asserting claims 9 and 10 of the '820 against Samsung's Galaxy XR headset (dToF), with no outcome yet.
• No other case, IPR, PGR, or CAFC appeal specifically involving the '820 was found as of my sources' latest updates. If you need this confirmed to a legal certainty, the authoritative checks are PACER for E.D. Tex. 2:26-cv-00356 and the USPTO PTAB portal (PTABE2E / PTACTS), searching the literal string "12,588,820."

I'll start by verifying the structured data against live sources, then search for any PTAB activity that may not yet be indexed in the ODP.

The structured data and search results point to sibling patents. Let me confirm there's no proceeding on the ’820 itself and check for any 2026 filings.

Let me do one more targeted check for any proceeding naming the ’820 directly, and confirm the status of the closely-related ’475 proceedings.

PTAB Proceedings on US 12,588,820

Proceedings overview

Total AIA trial proceedings on US 12,588,820: zero. The canonical structured source (USPTO ODP API) returns no IPR, PGR, EPR, or CBM proceeding naming this patent — and my live web checks confirm no petition, institution decision, FWD, or appeal anywhere in the public record names 12,588,820 (read literally as 12,588,820; I did not substitute a look-alike). Breakdown by status is therefore: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution-denied.

Bottom line for a defendant: the patent has a clean PTAB slate — but that is not the same as "hardened." The ’820 issued only on 2026-03-31, so it is too new to have accumulated challenges. Every other patent in the Omni MedSci continuation family is under attack or has already been invalidated, and the ’820 sits on the same specification and shares its immediate parent (U.S. App. 18/438,144, now US 12,193,790) with patents that are in active PGRs. Treat the empty docket as a timing artifact, not a defensive asset — and note that because the challenged siblings are being treated as AIA patents, the ’820 window is likely a PGR window, not an IPR window (see Strategic summary).

Source-of-truth reconciliation: The structured ptAB block in this prompt (ODP ingest) and web search agree — no proceeding on the ’820. Search surfaced several Law360 headlines ("PTAB will no longer conduct an IPR into the validity of an Omni MedSci wearable device patent, now that there will be a separate post-grant review," dated 2026-05-19). Those refer to IPR2025-01253 / PGR2026-00003 on the ’790 — not the ’820. Do not let a vendor summary misattribute them to the ’820.

---

Proceedings on US 12,588,820

None. There is no proceeding to report. I am not going to invent a number to fill this section.

---

Closely-related family proceedings (context — NOT proceedings on the ’820)

Because the ’820 shares a specification, a claim-drafting template, and the immediate parent application with the patents below, these are the proceedings a defendant actually needs to read. They are the closest thing to a "PTAB record" for the ’820’s subject matter.

IPR2021-00453 — Apple Inc. v. Omni MedSci, Inc. (U.S. Patent 10,517,484)

• Type: Inter Partes Review
• Filed: 2021-01-22 (Petition, Paper 1)
• Status: All challenged claims held unpatentable across two FWDs (original + remand). On appeal, Fed. Cir. No. 25-1646.
• Judge panel: Per the record cited in later briefs, the Board’s panel issued Paper 22 (2022-08-03) and Paper 26 (2025-02-14). Panel names not confirmed in my sources — I won’t guess.
• Petition grounds: §103 obviousness; claims 1, 2, 7, 15–23 (Ground set 1) and claims 3–6, 8–14 (alternative ground on remand).
• Institution decision: Instituted 2021-08-06 (Paper 7).
• Final Written Decision: Paper 22 (2022-08-03) held claims 1, 2, 7, and 15–23 unpatentable but found Petitioner had not shown claims 3–6 and 8–14 unpatentable. On remand (Fed. Cir. No. 2023-1034, Apple Inc. v. Omni MedSci, 2024 WL 3084509 (2024-06-21), reversed-in-part and remanded), the Board held claims 3–6 and 8–14 unpatentable, Paper 26 (2025-02-14). Net effect: every claim of the ’484 has been held unpatentable by the Board.
• Appeal: Yes. Apple appealed the original partial-loss (claims 3–6, 8–14) → No. 2023-1034 (reversed & remanded 2024-06-21). Omni then appealed the remand FWD → No. 25-1646 (Fed. Cir.), still pending.
• Defensive value: This is the single most useful invalidity roadmap for the family. The ’484 independent claims recite the same "wearable device / light source / lens / detection system / output signal" architecture that appears in the ’820’s SUMMARY. If the ’484 is affirmed on the 25-1646 appeal, an argument that the ’820’s independent claims are invalid over the same art becomes substantially stronger.

IPR2019-00916 — Apple Inc. v. Omni MedSci, Inc. (U.S. Patent 9,651,533)

• Type: Inter Partes Review
• Filed: 2019-04-10 (Petition, Paper 1)
• Status: Claims invalidated; affirmed on appeal.
• Judge panel: Not confirmed in my sources.
• Petition grounds: §103; challenged claims 5, 7–10, 13, and 15–17.
• Institution decision: Instituted 2019-10-18 (Paper 16).
• Final Written Decision: Paper 39 (2020-10-14) held all challenged claims (5, 7–10, 13, 15–17) unpatentable as obvious.
• Appeal: Yes — Omni MedSci, Inc. v. Apple Inc., No. 21-1229, 2022 WL 2062168 (Fed. Cir. 2022-06-08) (affirmed, reported as a summary affirmance). Compare the related first appeal addressing assignment standing, No. 20-1715.
• Defensive value: Confirms the core "dark-signal subtraction to improve SNR" and "synchronize detector to light source" concepts — recited in the ’820’s SUMMARY independents — were found obvious on this record. Note, however, that the ’533 claims asserted in current litigation (11, 12) were not part of this FWD and therefore survive as untested.

PGR2026-00003 / IPR2025-01253 — WHOOP, Inc. (PGR) and Samsung et al. (IPR) v. Omni MedSci, Inc. (U.S. Patent 12,193,790)

• Type: PGR2026-00003 = Post-Grant Review; IPR2025-01253 = Inter Partes Review (parallel, same claim — ’790 claim 7)
• Filed: IPR2025-01253 filed 2025-08-05; WHOOP’s PGR2026-00003 filed 2025-10-13
• Status: IPR2025-01253 — institution GRANTED, then VACATED and DENIED (no trial). PGR2026-00003 — INSTITUTED. This is the "separate post-grant review" the May 2026 Law360 items describe.
• Judge panel: Director John A. Squires issued both the sua sponte Director Review order and the vacatur/denial. Board panel for the underlying ’790 IPR: not confirmed.
• Key procedural event (this is the important one): On 2026-02-03 the Board granted institution of IPR2025-01253 (Paper 19). On 2026-04-01 the Director granted institution in PGR2026-00003. On 2026-05-06 the Director initiated sua sponte Director Review (Paper 27), reasoning that a patent "cannot be both PGR eligible and IPR eligible at the same time" and that the ’790 is an AIA patent (PGR-eligible, not IPR-eligible). The resulting order vacated the IPR institution and denied IPR2025-01253, leaving the PGR as the operative vehicle.
• Settlement / termination: None.
• Appeal: Not applicable to the vacatur.
• Defensive value / direct read-through to the ’820: This is the most important related proceeding for the ’820. The ’820 and the ’790 share the same immediate parent application (18/438,144) and the same 2012 priority claim. The Director’s "current record" finding that the ’790 is an AIA patent strongly signals the same treatment for the ’820. Practical consequence: a challenge to the ’820 is likely a PGR filed within 9 months of the 2026-03-31 grant, not an IPR.

IPR2025-01254 / PGR2025-00063 — Samsung et al. v. Omni MedSci, Inc. (U.S. Patent 12,268,475)

• Type: IPR2025-01254 = Inter Partes Review; PGR2025-00063 = Post-Grant Review (parallel, belt-and-suspenders on the AIA-eligibility question)
• Filed: both 2025-08-05
• Status: Petitions filed; institution status per the latest Delaware opinion is that the PTO has "instituted IPR, EPR, or PGR proceedings with respect to all the asserted claims in each of the asserted patents" — but I could not confirm the specific institution paper/date for the ’475 in my sources. Treat as likely instituted / pending rather than confirmed.
• Petition grounds: §103 — Ground 1 over Lisogurski (U.S. 9,241,676) + POSITA knowledge (claims 1, 8, 11–13); Ground 2 over Lisogurski + Tran (U.S. 8,108,036) (claims 5, 7); Ground 3 over Lisogurski + Tran + LeBoeuf (US 2010/0217102) (claims 6, 7). Petitioner expressly relied on collateral estoppel from the ’533 and ’484 FWDs on "identical and/or substantially identical" limitations.
• Defensive value / direct read-through to the ’820: The ’475 petition quotes a claim 1 that is nearly word-for-word the ’820’s SUMMARY apparatus claim — "biosensors … differential measurement … pulse rate monitoring and a blood flow measurement … plurality of light emitting diodes … one or more lenses … detection system … synchronized to the light source … plurality of detectors spatially separated … at least one analog to digital converter … increase the signal-to-noise ratio by increasing light intensity … generate a first signal responsive to light received while the light emitting diodes are off … second signal … comparing the first and second signal … determine … that the apparatus is being worn by the user." If the ’475 claims are held unpatentable over Lisogurski/Tran/LeBoeuf, that art maps directly onto the ’820’s independent claims.

IPR2025-01249 — Samsung et al. v. Omni MedSci, Inc. (U.S. Patent 9,055,868)

• Status: Institution DISCRETIONARILY DENIED — Paper 16 (2025-12-01). Google/PTAB sources cite this as the one Samsung petition the PTO declined. This is the family’s only defensive win for Omni at the institution stage.

Active/pending Samsung family petitions (context)

• IPR2025-01250 — ’533; institution GRANTED 2026-02-03 (Paper 19); scheduling order 2026-02-12. Panel includes APJs Grace Karaffa Obermann, John F. Horvath, Sharon Fenick (per Docket Alarm case header; verify on PTAB E2E).
• IPR2025-01251 — ’304; filed 2025-08-05.
• IPR2025-01252 — ’455; filed 2025-08-05.
• PGR2025-00064 — ’790 (Samsung); institution denied / mooted in favor of the IPR — which was itself later vacated (see above).
• WHOOP’s own challenges: IPR2025-01583 (’533 — Grounds over Lisogurski+Carlson, +Walker, +Tam) and IPR2025-01584 (’304).

---

Strategic summary

What is canceled vs. sustained vs. untested — for the ’820 specifically. Nothing is canceled, sustained, or tested. The ’820 has no PTAB adjudication of any claim. The patent’s own SUMMARY describes three independents (apparatus adapted to be worn; wearable device with on/off light source and processor; system for a worn device). I could not retrieve the issued claim set verbatim from the authoritative text supplied — the text stops mid-Detailed-Description — so any statement that "claims 1–N of the ’820 recite X" is an inference from the SUMMARY, not a verbatim reading. Someone relying on the claims must pull the printed patent from PatentCenter/PTAB E2E before building a chart.

Estoppel landscape. There is no §315(e)(2) estoppel arising from any proceeding on the ’820, because no such proceeding exists. Petitioners in the sibling IPRs (Apple in -00453/-00916; Samsung/Fossil/Oura/OnePlus in the 2025 series; WHOOP in -01583/-01584) are estopped only as to the patents they actually challenged. That leaves all prior-art grounds against the ’820 open to a defendant today — subject to one practical caveat: Omni is actively asserting collateral estoppel against itself in the family (the ’790 and ’475 petitions argue that the ’533/’484 FWDs, which found "identical or substantially identical" limitations obvious, preclude relitigation). If those FWDs are affirmed, expect Omni to face — and not be able to resist — the same reasoning on the ’820’s overlapping limitations.

Pattern signals. Clear and strong:

1. Every challenged claim in the ’533/’484 line has fallen (’533: claims 5, 7–10, 13, 15–17; ’484: all claims across the two FWDs). The only family claim sets that survived are the ones never challenged (e.g., ’533 claims 11–12) — i.e., Omni has won by drafting around the FWDs, not by defending claims.
2. Omni is a serial appellant, not a settler: it appealed the ’484 remand FWD (No. 25-1646), and the family has generated at least three Fed. Cir. appeals (20-1715, 21-1229, 23-1034).
3. No defensive aggregator (Unified Patents) appears in the chain; the petitioners are direct market competitors (Apple, Samsung, Fossil, Oura, OnePlus, WHOOP).
4. The AIA-eligibility fight is now the family’s central procedural battleground. The Director has now ruled (IPR2025-01253 vacatur / PGR2026-00003 institution) that the ’790 is an AIA patent — meaning challengers must litigate in PGR, not IPR, and the decision has already produced a split result (Karaffa Obermann panel granted the ’533 IPR the same day the ’790 IPR was vacated).

---

Recommended next steps

Because there is no PTAB activity on the ’820, the clock — not the docket — is the actionable fact.

1. Assume PGR, and assume the clock is running. The ’820 granted 2026-03-31. If it is an AIA patent (highly likely given the ’790 Director ruling on the shared parent 18/438,144), a PGR petition must be filed within 9 months of grant — i.e., on or before roughly 2026-12-30. After that, only IPR (and only if the Director’s "cannot be both" reasoning leaves an IPR path) or district-court invalidity remain. Do not wait for a Rule 16 conference. Confirm the exact PGR deadline against the patent’s grant date on PTAB E2E / PatentCenter.
2. Reuse the ’475 and ’484 records. The IPR2025-01254 / PGR2025-00063 petitions against the ’475 quote claim language that mirrors the ’820’s SUMMARY apparatus claim almost verbatim, keyed to Lisogurski (U.S. 9,241,676), Tran (U.S. 8,108,036), and LeBoeuf (US 2010/0217102). Pull those petitions and the corresponding expert declarations (Anthony, Mercier) as your starting §103 package for the ’820. (Ptacts petition documents: IPR2025-01254 petition exhibit set and the ’475 petition text reproduced in IPR2025-01250 Ex. 1096.)
3. Track the two Fed. Cir. appeals that control the family’s estoppel story. Apple v. Omni, No. 25-1646 (’484 remand FWD — all claims unpatentable) and the earlier No. 23-1034 decision. If the remand FWD is affirmed, the invalidity theory for the ’820’s overlapping limitations is materially de-risked. Docket: CourtListener / Fed. Cir. docket.
4. Watch the ’475 IPR/PGR outcome. It is the closest structural analog to the ’820 and will be the first adjudication of the "biosensor + differential measurement + LED + on/off SNR subtraction + wearer-detection" claim combination. Its FWD (due ~12 months from institution) is effectively a preview of how the Board will read the ’820’s independents.
5. Confirm nothing has been filed since ingest. The ODP block is authoritative but lags. Before relying on "zero proceedings," run a fresh PTAB E2E search for patent number 12,588,820 and a PatentCenter "Patent Trial Proceedings" tab check — the ’820’s 9-month PGR window makes a late-2026 filing entirely plausible.

One correction to carry forward: the previously generated "Patent summary" section flagged Law360 reports of the PTAB declining an IPR in favor of a PGR "on an Omni MedSci wearable-device patent." That reporting concerns the ’790 (IPR2025-01253 vacated; PGR2026-00003 instituted) — I could find no report tying it to the ’820. The two should not be conflated. Neither does the prior summary’s assertion that "Omni MedSci’s campaign pleadings assert the sibling ’475, ’790, ’484, ’304, ’455, ’533, and ’868 patents" extend to the ’820: my checks found no pleading asserting the ’820 in any of the identified suits.

I'll research the assignment record for US 12,588,820, starting with the sources I can query directly.

Let me dig deeper into the specific assignment records and correspondent details.

Let me search for the specific assignment reel/frame records and correspondent details.

Assignment Record Reconstruction — US 12,588,820 B2

Research limitation — stated up front (do not skip): I could not open the USPTO Assignment Center record set (assignmentcenter.uspto.gov / assignment.uspto.gov) directly in this session. Everything below is reconstructed from (a) the authoritative patent text you supplied, (b) Google Patents legal-events entries embedded therein, (c) the Federal Circuit's Omni MedSci v. Apple record and the Federal Circuit's own statement of the 2013 recordings, and (d) USPTO prosecution papers produced as PTAB exhibits (PGR2025-00063 / IPR2025-01249–01253). I do not have reel/frame numbers for any assignment in this chain, and I am not going to invent them. Where the field is unknown I say UNRETRIEVED. Treat every reel/frame placeholder below as an open item for direct Assignment Center verification.

---

Inventors

Pattern notes:

• Single-inventor family. Every application in this continuation chain — including 18/899,012 — names only Mohammed N. Islam. There is no co-inventor cohort, so the classic "all inventors departed within 12 months" fire-sale tell is inapplicable/not present.
• Employer-ownership conflict, litigated and resolved in favor of the applicant. This is the headline "unusual pattern." UM's Office of Technology Transfer asserted in 2013 that UM, not Omni, owned the inventions; Dr. Islam filed an Invention Report but did not pursue UM's internal appeal. Apple later moved to dismiss Omni's 2018 E.D. Tex. suit for lack of standing on the theory that UM Bylaw 3.10 was a present automatic assignment. The Federal Circuit held the bylaw language ("shall be the property of") was at most a promise of a future assignment, that Dr. Islam retained title, and that his December 17, 2013 assignment to Omni was valid (Omni MedSci, Inc. v. Apple Inc., 7 F.4th 1148 (Fed. Cir. Aug. 2, 2021); Newman, J., dissenting). This ownership cloud is the single most material diligence fact on the chain — it has already been litigated once and UM persists in its position (it briefed the appeal as a third party).

---

Original assignee

Omni MedSci, Inc., Ann Arbor, MI (per the patent front page and the 2024-09-27 recording).

• Entity type: Inventor-founded, small-cap private company formed by Dr. Islam in 2012 during his UM leave ("a new Biomedical Laser Company" per his own contemporaneous correspondence, J.A. 603).
• Primary line of business: Development and monetization of near-infrared / short-wave-infrared laser and optical-sensing intellectual property (supercontinuum fiber sources, LED/laser-diode wearables, time-of-flight sensing, camera/AI sensing). Its patent filings since 2023 have expanded into time-of-flight and generative-AI camera subject matter.
• Did it ship a product embodying the claims? No evidence found. I found no commercial product, no FDA clearance, no revenue disclosure, and no 10-K/8-K (Omni is private and does not file with the SEC — so the SEC cross-reference requested in the brief yields nothing). The company's observable activity is patent prosecution plus patent assertion. I flag this as unclear-to-negative, not as a proven absence.
• Current status: Operating as a going concern; not acquired, not dissolved, and no bankruptcy filing found. It is an active plaintiff in multiple districts as of 2025–2026.
• Real party in interest: Dr. Islam is the founder, the sole named inventor, the assignor on every record, and (per the Federal Circuit's 2021 background) the person who personally corresponded with Apple engineers in 2016–2017 before asserting.

---

Assignment timeline

Important: Google Patents' legal-events panel for US 12,588,820 shows exactly one post-filing reassignment entry. There are no security agreements, no releases, no corrections, and no name-change conveyances surfaced in any source I could reach. The 2013 recordings described by the Federal Circuit are recordings against the ancestor applications, not against the '820 itself (which did not exist until 2024).

1. 2013-12-17 (executed) / recorded December 2013 — Reel UNRETRIEVED / Frame UNRETRIEVED

• Conveyance: Assignment (Assignment of Assignor's Interest)
• Assignor: Mohammed N. Islam
• Assignee: Omni MedSci, Inc.
• Correspondent: UNRETRIEVED (not exposed in any source I could reach)
• Context: Founder-to-company assignment of the parent application family rooted in the Dec. 31, 2012 provisionals and the Dec. 17, 2013 PCT/US2013/075700, PCT/US2013/075736, PCT/US2013/075767 filings. The Federal Circuit expressly found: "After those applications issued as patents, Dr. Islam assigned the patent rights to Omni on December 17, 2013 and recorded their assignment with the Patent and Trademark Office." This is the founding record of the chain, not a transfer-to-asserter.
• Scope caveat: If this 2013 instrument recited "all continuations, divisionals and reissues," it may reach the '820 as a continuation. I cannot confirm scope without the instrument. Do not assume either way.

2. 2024-09-27 (executed/recorded) — Reel UNRETRIEVED / Frame UNRETRIEVED

• Conveyance: Assignment — Google Patents renders it verbatim as: "OMNI MEDSCI, INC. — ASSIGNMENT OF ASSIGNOR'S INTEREST — Assignors: ISLAM, MOHAMMED N."
• Assignor: Mohammed N. Islam
• Assignee: OMNI MEDSCI, INC.
• Correspondent: UNRETRIEVED. Likely flag (not a finding): the entire family is prosecuted under Customer No. 109543 — Brooks Kushman P.C., most recently at 150 W 2nd St., Suite 400N, Royal Oak, MI 48067 (previously 1000 Town Center, 22nd Floor, Southfield, MI 48075). Named practitioners of record across the family include Andrew B. Turner (Reg. No. 63,121), Kevin J. Heinl, Thomas A. Lewry, John S. LeRoy, John M. Halan, and Christopher C. Smith. Whether the assignment-cover-sheet correspondent is Brooks Kushman or someone else is not established by my sources.
• Context: Internal inventor-to-company recordation, i.e., a clean-standing formality executed on the filing date of application 18/899,012. This keeps the record of title current on the new continuation rather than reflecting any change of beneficial ownership.

3. No further transfers

No assignment out of Omni MedSci appears anywhere in the record — not to a holding LLC, not to a fund, not to a defensive aggregator. Per the "if the Assignment Center has no [further] records" rule: the original assignee still owns the patent.

Cross-references checked

• Google Patents legal events: two entries only — 2024-09-27 "Assigned to OMNI MEDSCI, INC." and 2024-09-27 "Application filed by Omni Medsci Inc." (plus 2024-10-25 priority link to 18/927,698, now US 12,268,475).
• USPTO PEDS/ODP & patent file histories (as produced in PTAB exhibits): confirm applicant of record OMNI MEDSCI, INC. and sole inventor Islam; no ownership change reflected in prosecution papers.
• SEC EDGAR: no filings exist — Omni MedSci is private. Requested cross-reference returns nil, which is itself expected and not a finding.
• RPX / Unified Patents asserter directories: Omni MedSci appears as a patent owner/plaintiff in Unified Patents litigation data and in RPX litigation documents (did not match the enumerated third-party NPE lists — see below).

---

Timeline diagram

timeline
    title Ownership of US 12588820
    2012 : Islam files provisionals while on UM leave
         : Islam founds Omni MedSci Inc
    2013 : Nonprovisional apps filed after return to UM
         : Islam assigns rights to Omni MedSci
         : Assignment recorded at USPTO
    2013 : UM technology transfer office disputes ownership
    2018 : Omni MedSci sues Apple on family patents
    2021 : Federal Circuit affirms Omni MedSci ownership
    2024 : App 18 899 012 filed by Omni MedSci
         : Islam assignment to Omni MedSci recorded
    2025 : Family PTAB challenges filed by Samsung group
    2026 : Patent US 12588820 B2 granted
         : Omni MedSci sues Samsung in E D Tex

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. The patent has never left Omni MedSci, Inc. There is no transfer to an "IP / Licensing / Holdings / Ventures" vehicle, no registered-agent-service address of record, and no single-purpose transferee LLC. The 2024 record is an inventor→operating-company instrument, not an operating-company→shell instrument. (Reel/frame UNRETRIEVED; conclusion rests on Google Patents legal events + Federal Circuit record.)

2. Known asserter in the chain — NOT PRESENT (as to the enumerated lists); UNCLEAR (high-activity plaintiff). Neither assignee — Islam or Omni MedSci — matches Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg entity. However, Unified Patents litigation data flags this family as subject to a coordinated defendant wave (Samsung + Samsung Electronics America + Fossil Group + Fossil Stores I + Fossil Partners LP + Oura Health + OnePlus, in IPR2025-01249–01253 and PGR2025-00063), which is the signature of a serial asserter even when the asserter is the original owner.

3. Repeat correspondent across the chain — PRESENT, but weak/attributionally benign. A single firm, Brooks Kushman P.C. (Customer No. 109543), is the correspondent of record on every application in this family, with Andrew B. Turner (Reg. No. 63,121) repeatedly signing IDS submissions across at least 17/078,771, 16/669,794, and 18/438,144, and Kevin J. Heinl listed as filer. Caveat per your own instruction: a single firm doing both NPE and operating-company patent work is not itself a finding, and here the recurrence is explained by a continuous prosecution relationship with a single inventor-founded client — not by a lawyer recycling a stable of anonymous shell LLCs. I record it as present-but-unpersuasive; the assignment-cover-sheet correspondent is UNRETRIEVED, so I cannot close this out.

4. Cascading transfers — NOT PRESENT. There is no chain of consecutive assignments through serial LLCs. One inventor→company record in 2013 (ancestors), one inventor→company record in 2024 (this app), and nothing else. Well outside the <24-month cascade pattern.

5. Pre-litigation transfer — NOT PRESENT. The 2024-09-27 recording postdates every Omni MedSci suit on the older asserted patents (Apple 2018; Apple 2020; Whoop 2025) and anticipates the 2026 Samsung/Fossil action by roughly 16 months, but it is a formality between the same inventor and the same long-standing owner, not a transfer to an assertion vehicle. Omni MedSci has held the family since 2013, so no standing was manufactured.

6. Bankruptcy fire-sale — NOT PRESENT. No Chapter 7/11 proceeding involving Omni MedSci or Islam was found; no court-supervised §365(n) sale.

7. Privateering — NOT PRESENT. No operating company transferred these patents to Omni MedSci. The flow is the reverse of privateering: an individual inventor's company asserting against large manufacturers who are not its patent source.

8. Defensive aggregator — NOT PRESENT. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. Note the irony: Unified Patents appears on the challenger side (its litigation data links the E.D. Tex. 2:26-cv-00356 record), i.e., the patent is a target, not a defensive acquisition.

Additional signal not in your list, but material — ownership-cloud / standing risk: PRESENT. UM's assertion of ownership was fully litigated (E.D. Tex. → N.D. Cal. → Fed. Cir. 2020-1715/-1716). Although Omni prevailed, Judge Newman's dissent ("overturns decades of unchallenged understanding") and UM's amicus position mean any acquirer inherits a non-trivial § 261/standing diligence item. See also Seagate (CMU), which cited Omni MedSci to knock out a license — the decision has become governing law that cuts both ways.

---

Verdict

NPE — moderate confidence.

Justification (2–3 sentences): The chain terminates at Omni MedSci, Inc., which has held the family continuously since the December 17, 2013 inventor assignment and the September 27, 2024 inventor assignment for application 18/899,012 — there is no shell transfer, no cascade, no bankruptcy, and no aggregator anywhere in the record, so this is not a classic patent-troll ownership chain. What pushes it into the NPE column is conduct rather than structure: an inventor-founded, privately held company with no evidence of any commercially shipped product embodying the claims is simultaneously suing Apple, Samsung, Fossil, Oura, OnePlus and Whoop in E.D. Tex., W.D. Tex. and D. Del. — a monetization-only footprint — and has triggered a coordinated multi-defendant PTAB wave (IPR2025-01249–01253, PGR2025-00063). Confidence is capped at moderate because the "no product" finding is an absence of evidence rather than affirmative proof, and because the assignee is the original owner and real-party-in-interest inventor, which is structurally unlike a shell-LLC asserter.

Verify directly (the two records I could not pull):

• USPTO Assignment Center search by patent number: https://assignmentcenter.uspto.gov/
• USPTO legacy Assignment Search: https://assignment.uspto.gov/patent/index.html

Specifically request/pull on that search: the reel/frame for both the 2013-12-17 and 2024-09-27 instruments, the cover-sheet correspondent on each, and the "assigns all" scope language of the 2013 instrument (to determine whether it reaches US 18/899,012 as a continuation). Those three data points are the only gaps in this reconstruction.

I'll research this. Let me first try to retrieve the "References Cited" / patent-citation record for US 12,588,820 specifically — the authoritative text I was given stops before the claims and does not include the front-page citation list.

I have the patent confirmed on Google Patents, but I still need the front-page citation list. Let me try to pull the granted-patent PDF and the Google Patents citation panels directly.

Prior Art for U.S. Patent No. 12,588,820 — Analysis

Read this before the tables. I searched for the citation record for the literal number 12,588,820. I have to be straight with you about what that search did and did not return, because the answer materially depends on it.

---

0. Retrieval status (do not skip)

I could not retrieve the front-page "(56) References Cited" list of US 12,588,820. Specifically:

• The authoritative patent text I was supplied (the Google Patents mirror of the USPTO record) stops mid-Detailed-Description, before the claims and before the front-page citation block. There is no "(56) References Cited" section in it.
• I could not open USPTO PatentCenter or the granted-patent PDF (patentimages.../US12588820.pdf) in this session.
• My web searches for "US12588820" + "References Cited" returned no front-page citation list for the ’820. What they returned instead were (a) Google Patents "family/citation graph" tables in which US 12,588,820 appears on other patents' pages, and (b) PTAB district-court exhibits listing the references asserted against the sibling patents — not the ’820's own examiner citations.

Therefore: the reference set below is NOT "the citations printed on the face of the ’820." I am not going to invent citation numbers to fill that table. What I can give you — and what is actually more probative for an invalidity analysis — is the art of record against the same-specification, same-title, same-immediate-family claims, which the petitions themselves treat as mapping limitation-for-limitation onto the ’820. I flag every inference as an inference.

Claim-text caveat (carried forward). I still do not have the issued ’820 claim set verbatim. The closest verified proxy I found is US 12,268,475 B2, which carries the identical title ("Wearable device for differential measurement on pulse rate and blood flow") and, per its Google Patents listing, contains:

• a wearable-device claim with the on/off light source and processor steps (i)–(iv) that match the ’820 SUMMARY almost word-for-word, and
• a system claim reciting the multiple-LED source, synchronized detection system, spatially separated detectors, ADC, and first/second-signal SNR comparison that likewise matches the ’820 SUMMARY.

That materially corroborates the earlier section's point that the ’475 and ’820 are near-twins, and it means the ’475 record is the best available stand-in for the ’820.

---

1. What I positively confirmed about the ’820

A contradiction with the earlier "Patent summary" section must be flagged: the litigation section reports that the complaint characterizes the ’820's asserted claims 9–10 as covering laser diodes, a photodiode array, first and second cameras, and combined time-of-flight/image data — i.e., a dToF thread, not an LED/biosensor thread. The SUMMARY-based characterization is the LED/biosensor thread. Both can be true (the ’475 listing itself contains both a light-source/lens/detector claim family and a later claim reciting "a laser diode configured to be pulsed … Bragg reflectors … beam splitter … time-of-flight"). Which thread claims 9–10 actually fall in is unverified. This matters because it changes the relevant art.

---

2. Patent-document prior art ("patent citations")

The following are the U.S. patent documents of record against the ’820's family, drawn from the PTAB exhibit lists in IPR2025-01250 / -01252 / -01254 and PGR2025-00063/00064 and from Apple's 2018 invalidity contentions. Dates and short names are as they appear in those records; titles I could not independently verify are marked "(title not verified)."

Also appearing in the ’820's Google Patents citation/family graph (role ambiguous — I could not tell whether these are "Citations," "Cited By," or "Family Cites Families," so I do not assert them as examiner citations):

• U.S. Patent No. 7,259,906 (Islam, "System and method for voice control of medical devices") — same inventor/assignee family.
• U.S. Patent Pub. No. 2009/0244288 A1 ("Imaging apparatus and lighting apparatus for imaging").
• Omni-family members US 9,500,635, WO 2014/143276, US 12,484,787 — these are the ’820's family, not prior art.

Critically: several of these references (Lisogurski, Carlson) post-date or straddle the Dec. 31, 2012 priority date, and US 9,596,990 (Nov. 6, 2013) post-dates it. Their availability as §102 art depends entirely on the effective-filing-date question in §5.

---

3. Non-patent literature of record (printed publications)

These were listed in the Omni-family petition exhibit sets (IPR2025-01250 Ex. 1035–1036; IPR2021-00453 Ex. set; Apple 2018 contentions):

---

4. §102 vs. §103 — the honest answer

On the record as it stands, I found no single reference that §102-anticipates an ’820 independent claim. Every ground pleaded in the family is a §103 combination (Lisogurski + Carlson, Lisogurski + Carlson + Mannheimer, Lisogurski + Carlson + Park ± Mannheimer). The strongest §102 candidates are narrow:

• Carlson (2005/0049468) — arguable single-reference §102 coverage of individual limitations (LEDs, lenses, ambient/environmental-light suppression, increased optical power → SNR). It does not alone disclose the "third SNR greater than the first and second SNR" or the wearer-detection limitation.
• Li et al. (June 2012) — arguable single-reference §102 coverage of the baseline-subtraction limitation; not the full combination.
• Mio Alpha — arguable §102 as to the preamble + LED-light-source elements only.

The limitations that impose the real §102/§103 burden (and that I would build the analysis around, pending the verbatim claims) are:

1. "a third output signal having a third SNR that is greater than the first SNR and greater than the second SNR" — the References' art shows improvement, not a third signal greater than both inputs; this is the classic §103 aggregation argument.
2. "determine … that the wearable device is being worn by the user" — a wearer-classification step; I found no single reference in the family record that discloses it.
3. "differential measurement" of pulse rate and blood flow as the two named parameters.
4. "at least one analog-to-digital converter is coupled to at least one of the plurality of detectors" combined with spatially separated detectors and source-synchronized detection.

If the asserted claims 9–10 are in fact the dToF/camera thread (per the complaint characterization — see §1), then the relevant art shifts entirely to time-of-flight/SPAD/laser-diode-with-Bragg-reflector art, and the family's own US 10,660,526 ("Near-infrared time-of-flight imaging using laser diodes with Bragg reflectors") and the Samsung laser-autofocus/ToF prior art would be the operative references. I could not confirm which is correct.

---

5. Statutory-date complication (this is the live issue, not a footnote)

The ’820 claims 2012-12-31 priority but was filed 2024-09-27. Which statute governs determines which of the above is even prior art:

• If pre-AIA §102(a)/(b)/(e) applies (2012 effective filing date): any reference published after 2012-12-31 drops out. That eliminates Schubert (2014), Park (Nov. 6, 2013) as §102(a) art, and Mio Alpha (2013–2014) becomes date-sensitive. Lisogurski (2012-05-31) and Carlson (2005) survive.
• If AIA §102(a)(1)/(a)(2) applies: the prior sections record that the Director held the shared-parent ’790 to be an AIA patent (Order vacating IPR2025-01253, 2026-05-06, on the ground that a patent "cannot be both PGR eligible and IPR eligible"). If the same holding attaches to the ’820, AIA §102(a)(2) makes U.S. patent documents with earlier effective filing dates available regardless of publication date — widening the art set (and making Park-type references available).

Practical consequence: the same reference can be §102 art under one regime and not the other, so pin the AIA/pre-AIA determination before committing to any anticipation theory. This is the same battleground the family is already litigating.

---

6. Flagged contradictions and open items

1. Claim-set subject matter. Earlier sections describe the ’820 as an LED/biosensor patent; the litigation section reports the complaint characterizes asserted claims 9–10 as laser-diode / photodiode-array / camera / time-of-flight. Unresolved. I did not pick a winner.
2. I could not produce the ’820's own (56) citation list. Anything presented as "the citations of US 12,588,820" by a third party should be verified against PatentCenter before reliance.
3. No PTAB/IPR/PGR or CAFC proceeding names the ’820 — consistent with the prior sections, and consistent with the patent being ~4 weeks old as of the stated date of April 26, 2026 (grant 2026-03-31). Sources I retrieved include events through 2026, so the date discrepancy noted in the litigation section persists; I did not suppress post-April material.
4. References #1, #4, #5 titles are unverified. Short names and dates come from PTAB exhibit lists; I did not independently confirm the patents' titles.

---

7. Recommended verification path (to close the gap)

1. Pull the granted PDF for 12,588,820 and read the (56) References Cited block directly — that is the literal answer to your question and it is the one item I could not obtain.
2. Cross-check the file wrapper for application 18/899,012 in PatentCenter for the examiner's PTO-892/PTOL-326 forms — these list the examiner-cited references, which differ from applicant-submitted IDS art.
3. Pull the IPR2025-01254 / PGR2025-00063 ’475 petitions and their exhibits (Lisogurski, Carlson, Tran US 8,108,036, LeBoeuf US 2010/0217102) — because the ’475 is the ’820's nearest structural twin, those petitions are the most efficient ready-made §103 package.
4. Confirm whether the Director's AIA holding on the ’790 extends to the ’820 before fixing the §102 statutory basis.

Bottom line: the most relevant prior art for US 12,588,820 on the available record is Lisogurski (US 9,241,676), Carlson (US 2005/0049468), Mannheimer (US 5,746,206), Park (US 9,596,990), Tam (US 7,029,628), and the Mio Alpha publication set, supported by the Li et al. (2012), Rhee et al. (2001), Pelaez (2007), Luo et al. (2007) and Asada printed publications — deployed in §103 combinations, with no confirmed single-reference §102 anticipation of an ’820 independent claim. The examiner's own citation list for the ’820 remains UNRETRIEVED.

I'll ground this analysis in the actual prior-art record for this family — the references the Board and petitioners have already used against the sibling patents, plus the '820's own cited art. Let me pull those.

Now let me check the '820's own face-cited prior art and any challenge mounted specifically against it.

Let me nail down the '820's own claim set and its face-cited references.

Obviousness Analysis — U.S. Patent No. 12,588,820 B2 under 35 U.S.C. § 103

Date caveat: your task header says April 26, 2026; the operating date for this analysis is September 28, 2026, and the sources retrieved include material dated through mid-2026. As in the earlier sections, I report everything found and flag the discrepancy rather than suppressing post-April events.

Building on the prior sections: this section assumes the bibliographic, litigation, PTAB, and assignment findings already generated. I do not repeat them. It builds directly on the PTAB section's finding that the closest structural analog to the '820 is the '475 (U.S. 12,268,475) — same title, same specification, same claim template — and on the '533/484 Final Written Decisions.

---

0. Threshold caveat and one flagged contradiction

I still cannot retrieve the issued claim set of the '820 verbatim. The authoritative text supplied stops mid-Detailed-Description. Everything below is therefore grounded in three things, in descending order of reliability:

1. The SUMMARY section of the '820 itself (authoritative, supplied), which recites the independent claims nearly word-for-word;
2. The verbatim claim text of U.S. 12,268,475 claim 1 and claim 8 as reproduced in the Samsung/WHOOP petitions (IPR2025‑01254 / PGR2025‑00063), which is character-identical to the '820's SUMMARY claim language — this is the single most useful corroboration available;
3. The '533 and '484 FWDs, which adjudicated the same limitations.

Contradiction flagged from the prior "Litigation summary" section: that section states the E.D. Tex. 2:26‑cv‑00356 complaint asserts "at least claims 9 and 10" of the '820 and describes the '820 as directed to "laser diodes, a photodiode array, first and second cameras, and combined time-of-flight/image data." That description does not match the '820's own abstract, title, or SUMMARY (which are pure LED/photoplethysmography), and it is a much better fit for a different Omni continuation. I could not verify claims 9–10 or the ToF characterization. I analyze the SUMMARY claim set (high confidence) and add a conditional analysis of a possible ToF/camera claim set in §9. Do not treat §9 as settled.

---

1. Governing law and framework

Because the '820 claims priority to December 31, 2012 provisionals through a chain including filings before and after March 16, 2013, the applicable statute turns on the AIA effective-filing-date question — the exact issue the Director decided for the sibling '790 in the IPR2025‑01253 vacatur / PGR2026‑00003 institution. Substantively this changes little for §103: pre-AIA §103(a) and AIA §103 are analyzed identically for the combinations below, and each reference independently qualifies as prior art under either regime. The one place it matters is the scope of "prior art" for Lisogurski (see §2.1).

The analysis follows the Graham v. John Deere factors: (1) scope and content of the prior art; (2) differences between the prior art and the claims; (3) the level of ordinary skill; (4) secondary considerations. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), governs motivation and the predictability inquiry.

Level of ordinary skill (POSITA). In the family's IPR record the parties framed the POSITA as a person with at least a bachelor's degree in electrical engineering, biomedical engineering, physics, or a related field, plus roughly 2–4 years of experience designing optical physiological sensors (or equivalent education/experience), with working knowledge of pulse oximetry, LED drive circuitry, photodetection, analog front-ends, and signal processing. I am not reproducing this verbatim — the '820's own record has not been made and I did not retrieve the '820-specific declaration. A POSITA of this description would know LEDs, lenses, photodiode arrays, ADCs, dark subtraction, and ambulatory motion/ambient-light artifact as routine design vocabulary. Note that in the family proceedings neither party materially disputed the POSITA definition, which reduces its importance as a battleground.

---

2. Scope and content of the prior art (the "Prior Art" record)

The '820's Google Patents page carries prior-art keywords "light, fiber, infrared, laser, processor" and a prior-art date of 2012‑12‑31 — i.e., the face of the patent itself concedes a 2012 optical-sensor art field. The operative references are those already deployed against this specification in the family:

2.1 Lisogurski as the primary reference — what it actually teaches

The Board and the Federal Circuit have already resolved Lisogurski's content and prior-art status, so it is settled ground rather than argument:

• Wearable sensor. Sensor 102/312 is a pulse oximeter "placed at a site on a patient," expressly including "the wrist to monitor radial artery pulsatile flow," as well as fingertip, toe, earlobe. '484 FWD 23, 26–27.
• Multiple LEDs. Light source 130/316 includes "a red LED and an IR LED," the IR LED emitting "between about 800 and 1000 nm." Id. at 10:52–58.
• Spatially separated detectors. Sensor "may include multiple light sources and detectors [that] may be spaced apart." 17:42–45.
• Detection system synchronized to the light source. Front-end processing circuitry 150 "synchronize[s] the operation of an analog-to-digital converter and a demultiplexer with the light drive signal." 11:41–46.
• ADC. ADC 154 in front-end circuitry 150. 11:43–46; 14:46–55.
• SNR by increasing light intensity. "The system may increase the brightness of the light sources in response to [any] noise to improve the signal-to-noise ratio." 9:50–52.
• Dark subtraction (the "first signal / second signal / compare" limitation). "Signal modification techniques, such as its dark subtraction technique … may be employed to reduce the effect of ambient light … and thereby reduce the contribution of the noise component to the converted digital signal." 14:46–55. Petitioners map this directly onto [1.k]–[1.m]. '533 FWD accepted this mapping; Fed. Cir. affirmed.
• Determining the device is being worn. "[T]he system may detect a signal indicative of a system error such as … a probe-off signal"/"physiologically impossible value." 36:66–37:2. Petitioners map this onto [1.n].
• Lens (or obvious variant). The '533 FWD held Lisogurski's LED encapsulant lens met "one or more lenses," and alternatively that a POSITA would add Carlson's lenses. '533 FWD 35–37.

§102 date nuance worth preserving: Lisogurski published December 5, 2013 — after the '820's December 31, 2012 provisional. Its prior-art force therefore rests on its May 31, 2012 filing date under §102(e) (pre-AIA) or §102(a)(2) (AIA), not on its publication. The Board so held and the Federal Circuit affirmed; Omni did not dispute it in the '484 IPR. Do not expect this to be a viable attack on the '820 either, but it is the one place where the AIA/pre-AIA question could theoretically bite.

---

3. Claim 1 / the apparatus claim — element-by-element

The SUMMARY apparatus claim, corroborated verbatim by '475 claim 1, breaks down as follows:

Result: no limitation of claim 1 lacks a Lisogurski-based teaching. The only genuine question is the motivation to integrate — addressed in §5.

---

4. The other two independents

Claim 2 (wearable device; light source on/off; lens; detector; processor generating first/second/third output signals; third SNR > both; wearer determination). This is the same substance expressed as a processor-method claim. The "first output signal / second output signal / third output signal with improved SNR" is the dark-subtraction result recast in signal-processing terms: Lisogurski teaches that dark subtraction "reduce[s] the contribution of the noise component," i.e., improves SNR — exactly the "third SNR greater than the first and second" limitation. The Board accepted this precise mapping for '475 claim 8, whose text is identical to the '820's SUMMARY claim 2, and the Director instituted PGR2026‑00003 on the '790, which recites the same processor architecture. A POSITA would treat the "third signal has better SNR than either input" recitation as an inherent, expected consequence of differencing a signal against a dark reference —KSR's "predictable result of a known technique."

Claim 3 (system for a worn device; includes "physiological parameters change in response to stretching of a hand, movement of a finger, or movement of a thumb"). This is the wrist/hand-worn limitation. It is the weakest defensible limitation, because it recites a property of the subject, not of the device. Lisogurski expressly places the sensor on the wrist to monitor radial artery pulsatile flow; Cai et al. (2010) discloses a wrist-band wireless pulse oximeter; Rhee et al. (2001) discloses a finger-ring plethysmographic sensor; Patterson et al. (2011) addresses artifact reduction in low-power reflective PPG. A skilled artisan in 2012 knew that wrist/finger-mounted PPG signals vary with hand and finger motion, and that this is why motion-artifact rejection was a live design problem. Under KSR / In re Schreiber, a claim reciting a result that necessarily follows from placing the sensor where the art already puts it adds nothing patentable. Expect Omni to argue this limitation imports a structural requirement; expect a petitioner to respond with an indefiniteness/written-description squeeze under §112, which is exactly the maneuver Omni pre-empted by getting Petitioners to stipulate that no indefiniteness dispute remains (the December 2025 email exchange in IPR2025‑01250).

---

5. Motivation to combine — the KSR-satisfying reasons

This is where the family record is strongest for a challenger, and it is why I assess the §103 exposure as high. Five independent, mutually reinforcing rationales exist:

(a) The reference itself suggests the modification. Lisogurski expressly says "the functionality of some of the components may be combined in a single component" and "may be divided over multiple components." 15:66–16:9. The Board held this expressly supplies the motivation to relocate control/light-drive/front-end circuitry into the sensor. '533 FWD 23. This defeats In re Ratti-style "bodily incorporation" arguments.

(b) Same field, same problem, complementary solutions. Every reference is an optical physiological monitor using LEDs, directed at the identical problem: extracting a small pulsatile signal against ambient light and motion noise. Lisogurski and Carlson "teach complementary and combinable methods for increasing signal-to-noise in a wearable pulsoximeter in the presence of ambient light or sunlight." Mannheimer removes skin-layer interference by spaced detectors; Lisogurski itself teaches that "light is attenuated differently depending on the tissue," including skin. LeBoeuf's arc captures angularly scattered light.

(c) Industry-wide trend (documented in printed publications). The 2010–2012 literature — Swan's "Sensor Mania! The Internet of Things, Wearable Computing, Objective Metrics, and the Quantified Self 2.0" (2012); Cai et al. wrist-band wireless pulse oximeter (2010); Rhee et al. finger-ring PPG (2001); Li et al. wireless reflective pulse oximeter with digital baseline control (2012, IEEE TBioCAS); Patterson et al. artifact reduction (2011) — establishes that migrating laboratory pulse oximetry into consumer wrist-worn form factors was the recognized direction of the art. Where "market demand … design incentives … [and] the identified, predictable solutions" point the same way, KSR holds the combination obvious.

(d) Predictable arrangement of known elements. Lenses, ADCs, spaced photodetectors, and dark subtraction each perform the same function they were known to perform, yielding the expected improvement. KSR; Sakraida v. Ag Pro.

(e) The Examiner's own allowance rationale is vulnerable. The Notice of Allowance quoted in the '790 PGR petition shows the Examiner allowed over prior art only because it did not disclose "detection system … synchronized to the light source," "spatially separated detectors," "analog to digital converter … coupled to … detectors," "increase the SNR by increasing light intensity," and the on/off first-and-second-signal comparison. Lisogurski discloses every one of those phrases. That is a §325(d) inversion: the reason for allowance is the reason to institute.

---

6. Reasonable expectation of success

Both Lisogurski and Carlson (and Tran, LeBoeuf, Mannheimer, Park) disclose working optical physiological measurement systems using LEDs, and the techniques (adding lenses, pulsing at a carrier frequency, placing detectors in an arc, differencing a dark sample) are routine optical-engineering steps with predictable outcomes. There is no teaching away, no inoperability, and no criticality of any undisclosed parameter. The Board has already found a reasonable expectation of success for the Lisogurski↔Carlson lens combination ('533 FWD 37) and for the Lisogurski↔Tran AI combination.

---

7. Dependent-claim mappings (illustrative)

Note on the "voice input" / "touch screen" elements: these are the kind of laundry-list hardware recitations that KSR and In re Venner treat as obvious where the reference monitor is already described as a portable, battery-powered computing device with a touchscreen and user input.

---

8. Where the patent has any defensive value — and the secondary-considerations position

The family's only actual wins have been limitations, not claims:

1. "Configured to identify an object" — the '484 claims 3–6 and 8–14 initially survived the Board's FWD on this construction; Omni's patent-owner construction ("to recognize or establish an object as being a particular thing") carried the day at the FWD stage. But that is now moot: the Federal Circuit vacated and remanded (Apple v. Omni MedSci, No. 2023‑1034, 2024 WL 3084509 (Fed. Cir. June 21, 2024)), and on remand the Board held claims 3–6 and 8–14 unpatentable (Paper 26, Feb. 14, 2025). Omni's appeal is pending (No. 25‑1646).
2. The "increasing a pulse rate" limitation — Omni won this at the institution stage of the '533 IPR but lost it in the FWD; the Federal Circuit affirmed.

The '820's independents do not even recite the "identify an object" language. [1.n] recites determining "that the apparatus is being worn" — which is the "detect an object" genus (see '484 claim 16, which the Board found unpatentable even before the remand). That is a materially weaker position for Omni than claims 3/8 ever were, and claims 3/8 lost on remand anyway. The prior "PTAB challenges" section's caution that the empty docket is "a timing artifact, not a defensive asset" is confirmed by this claim-level read.

Secondary considerations. As of this analysis I found no evidence of objective indicia in the family record: no nexus-bearing commercial product (consistent with the assignment section's finding that Omni ships nothing embodying the claims), no unexpected-results data, no copying evidence, no licensing-due-to-validity. Crucially, Omni did not prevail on secondary considerations in the '533 or '484 IPRs — its wins were construction-based. The '475 petition devotes a specific section to pre-empting secondary considerations, which signals that Petitioners expect Omni to attempt them.

---

9. Conditional analysis: if the asserted claims really are ToF/camera claims

Flagged as unverified. If the 2:26‑cv‑00356 complaint in fact asserts '820 claims directed to laser diodes + photodiode array + first/second cameras + combined time-of-flight/image data (as the earlier litigation section reports), then the operative art changes. On the present record the most relevant disclosure is Omni's own family — e.g., the Islam continuation described in the IEEE patent-citation listing as "laser diodes (LDs) generating near-infrared wavelength light … the first receiver … synchronized to the pulsed light and configured to perform a time-of-flight measurement … an infrared camera … captures light while the second part is off, and while the second part is on … and differences the signals to generate an image." That is an Islam-authored disclosure of the very combination. Note the two consequences if so:

• It is prior art against the '820 on its own §102(a)(2)/§102(e) filing date, being an earlier-filed Omni application published before the '820 was filed (2024‑09‑27). An inventor's own earlier continuation is §102(e) art as to a later continuation in the same chain unless the '102(b)(2)(C)/§103(c) common-ownership exception applies — and here common ownership is same-inventor, same-assignee, which is precisely where that exception does apply.
• Therefore a ToF/camera claim set would need third-party ToF art (structured-light/dToF depth-sensing references from the 2010s) rather than the Omni family. I did not retrieve that art in this session and will not assert it exists. Treat §9 as a research lead, not a conclusion.

---

10. Bottom line

Assessment: the independent claims of US 12,588,820 are highly vulnerable to a §103 challenge, and the vulnerability is unusually well-documented because the Board has already adjudicated essentially every limitation, on this specification, over this art.

• Primary combination (Ground 1): Lisogurski alone + POSITA knowledge. Lisogurski teaches or renders obvious every element of claim 1 — wearable pulse-oximeter biosensor, multiple LEDs, lens, ADC-coupled spatially separated detectors, synchronization to the light drive signal, SNR improvement by increasing LED brightness, dark subtraction (the "first signal/second signal" comparison), and probe-off detection (the "being worn" determination). The only required "combination" is the relocation of front-end processing circuitry into the sensor — which Lisogurski expressly suggests.
• Ground 2: Lisogurski + Tran for the AI/processor and smartphone limitations.
• Ground 3: Lisogurski + LeBoeuf (detectors along an arc) and/or + Carlson (beam-shaping lenses; pulsed LEDs outside the ambient optical spectrum) and/or + Mannheimer (spatially differential skin-layer subtraction) for the dependent claims.
• Motivation: the reference's own express suggestion; identical field and identical SNR problem; a 2010–2012 documented industry migration to wrist/finger wearables; and predictable, same-function-same-result assembly of known elements — the full KSR catalogue.
• Expectation of success: high and already found by the Board for the Lisogurski↔Carlson and Lisogurski↔Tran sub-combinations.

The strongest remaining counterarguments for Omni are (i) the "being worn" limitation, if the Board adopts a narrow "identify an object"-style construction — but note the Federal Circuit already vacated that construction and the remand FWD killed claims 3–6/8–14; (ii) AIA/pre-AIA status and Lisogurski's §102(e)/(a)(2) availability; and (iii) the possibility that the actually-asserted claims are ToF/camera claims requiring a different art set (§9).

What I could not do, and what must be done before relying on this: (1) obtain the printed claim set of the '820 (PatentCenter / PTAB E2E, literal string "12,588,820") — every statement above about specific claim numbers is an inference; (2) confirm whether any PGR or IPR was filed against the '820 after the ODP ingest, given the ~2026‑12‑30 nine-month PGR window; (3) confirm the actual asserted claims in E.D. Tex. 2:26‑cv‑00356 (PACER), because if the earlier section's ToF/camera characterization is right, §9 governs and §3–§7 do not.

Sources relied on for the prior-art mappings:

• Lisogurski–Carlson–Tran–Mannheimer claim charts and the '533 FWD, as reproduced in the Apple IPR2021‑00453 petition: https://knobbemedical.com/wp-content/uploads/2021/08/Petition-for-Inter-Partes-Review.pdf
• '475 claim 1/claim 8 verbatim and the Lisogurski ground mapping (IPR2025‑01254 / PGR2025‑00063): https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1558168](/patent/1558168)/download-documents (and the mirrored petition text at https://www.docketalarm.com/cases/PTAB/IPR2025-01250/…/Exhibit-1096-Ex_1096___475_Petition_IPR2025_01254.pdf and https://ai-lab.exparte.com/case/ptab/IPR2025-01254/doc/1027)
• '790 PGR ground and the Examiner's allowance rationale: https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1558540](/patent/1558540)/download-documents
• Federal Circuit '484 decision (identify/detect limitations): Apple Inc. v. Omni MedSci, Inc., No. 2023‑1034 (June 21, 2024), reproduced at http://fedcircuitblog.com/wp-content/uploads/2024/07/Omni-Petition.pdf
• Director's IPR2025‑01253 vacatur (AIA/PGR-eligibility): https://www.uspto.gov/sites/default/files/documents/IPR2025-01253_Samsung_v_Omni_MedSci_DR_Decision.pdf
• Family NPL (Li 2012; Patterson 2011; Rhee 2001; Cai 2010; Schreiner 2010; Swan 2012) via the '475 references-cited listing: https://patents.justia.com/patent/12268475
• '820 bibliographic/prior-art metadata: https://patents.google.com/patent/US12588820/en