US 10900770

Patent Overview: US10900770B2

Title: Distance measuring device, imaging apparatus, moving device, robot device, and recording medium

Patent number: US10900770B2 | Application: US16/353,573 | Publication (grant): 2021-01-26

Assignee: Canon Inc. (assignment recorded to CANON KABUSHIKI KAISHA, 2019-06-06; inventor Kazuya Nobayashi)

Key dates: Priority date 2018-03-27; filed 2019-03-14; granted 2021-01-26; anticipated expiration listed as 2039-03-14. Status listed as Active.

Purpose: The invention concerns a distance measuring technique using an imaging plane phase difference method. A distance measuring device is described as mountable on a moving member — for example, a vehicle, a drone, or a robot — to measure distance between that moving member and its surroundings or obstacles in order to support action. A recognition process using an image is required, so an imaging apparatus performs the distance measurement. Cited Japanese Patent No. 5917207 is described as disclosing an imaging apparatus using an imaging plane phase difference method that can acquire both an image for recognition and a distance.

Plain-language scope: Two image data are acquired from an optical image formed by light fluxes passing through different pupil regions of an imaging optical system. A relative positional deviation ("parallax amount") between images is detected as in stereo imaging, and converted to a defocus amount via a predetermined conversion coefficient, which in turn converts to object distance using the imaging-forming relation. A composite image or one of the images can serve as the captured object image. The text identifies noise in the image as one error factor: with little received light per photoelectric conversion unit pixel, noise influence grows and distance reproducibility falls, so each pixel portion should desirably receive light in a wide wavelength band. Axial chromatic aberration of the imaging optical system is also noted as affecting distance information, though the excerpt is truncated at that sentence.

Source: https://patents.google.com/patent/[US10900770](/patent/US10900770)/en

Not established by the excerpt: The full text of the claims and their exact scope, the complete chromatic-aberration discussion, figure contents, and any specific embodiments (e.g., filter or pixel structures) are omitted. The excerpt also does not establish whether the "Active" legal status or the anticipated expiration date has been verified.

I'll search for litigation involving US patent 10900770 across patent litigation databases.

The first result concerns a different patent (10,790,703). Let me search more specifically.

None of those results concern US 10,900,770 specifically. Let me search more targeted litigation databases.

Litigation Search: US Patent No. 10,900,770

Summary Finding

I found no known litigation involving U.S. Patent No. 10,900,770.

Searches across patent-litigation sources (Unified Patents litigation portal, RPX Insight litigation documents, CourtListener/PACER-derived dockets, USPTO PTAB filings, and general web sources) returned no case in which US 10,900,770 was identified as a patent-in-suit, an asserted patent, or otherwise a subject of a filed action.

What the Searches Actually Returned (and why it isn't 10,900,770)

It is important to report these carefully, because several hits are near-miss numbers that must not be conflated with 10,900,770:

The only Canon-related patent suits surfaced involve unrelated Canon patent families (printer/toner cartridges, digital camera patents), and the only camera/lens-imaging Canon case found (Optimum Imaging Technologies LLC v. Canon Inc.) asserted U.S. Patents 7,612,805 and 8,451,339 — not 10,900,770.

Patent 10,900,770 — Bibliographic Context (from the authoritative patent text provided)

For completeness and to confirm the identity of the patent searched:

• Title: "Distance measuring device, imaging apparatus, moving device, robot device, and recording medium"
• Patent No.: US 10,900,770 B2
• Application No.: US 16/353,573
• Inventor: Kazuya Nobayashi
• Assignee: Canon Inc. (Canon Kabushiki Kaisha)
• Priority date: March 27, 2018
• Filing date: March 14, 2019
• Grant date: January 26, 2021
• Anticipated expiration: March 14, 2039
• Legal status: Active
• Subject matter: Imaging-plane phase-difference distance measurement using a divided optical filter (first/second spectral transmittance regions)

Conclusion and Caveats

• No district court litigation (D. Del., E.D. Tex., N.D. Cal., S.D. Cal., etc.), no CAFC appeal, and no PTAB proceeding (IPR/PGR) involving US 10,900,770 was identified.
• Caveat on completeness: This conclusion is based on the web sources reachable during this search session. A negative finding is inherently limited — public litigation databases can lag filings by days to weeks, and complaint-level patent number identification sometimes appears only in exhibits. To be certain, a direct docket query would be advisable (e.g., Docket Navigator, Lex Machina, or PACER full-text search for "10,900,770"), which I could not fully execute here.
• I did not auto-correct any patent number. Every number above is reported exactly as it appeared in the source results, and none matches 10,900,770.

If you would like, I can attempt additional targeted queries (e.g., by assignee Canon Inc. as plaintiff, or by inventor Nobayashi) to try to surface any case in which this specific patent may have been asserted.

I'll verify the structured data against live sources before reporting.

No hits. Let me run additional targeted verification queries.

Proceedings overview

Total AIA trial proceedings on US 10,900,770: zero. The canonical structured source in this prompt (USPTO Open Data Portal / PTAB proceedings on file) returns no IPR, PGR, or CBM proceeding naming US 10,900,770 as the challenged patent, and my live searches corroborate that (breakdown by status is therefore: active 0 / claims invalidated 0 / claims sustained 0 / settled 0 / institution denied 0).

Bottom-line defensive posture: there is no PTAB record to lean on — neither a shield nor a roadmap. This is not the "claims 1–5 have been canceled" scenario, and it is not the "patent survived two IPRs and is hardened" scenario either. It is the third, less-common posture: the patent has never been tested at the Board at all. Any defense must be built from scratch — no FWD to point to, no claim-construction ruling to borrow, no § 315(e)(2) estoppel to exploit, and no petitioner-side work product you can free-ride on.

---

Per-proceeding detail

None — no proceeding exists to analyze

• Type: N/A (no IPR, PGR, or CBM on file)
• Filed: N/A
• Status: N/A — the ODP structured block contains no AIA trial for this patent, and web searches returned no petition, institution decision, FWD, or termination naming 10,900,770.
• Judge panel: N/A
• Petition grounds: N/A
• Institution decision: N/A
• Final Written Decision: N/A
• Settlement / termination: N/A
• Appeal: N/A — consequently there is no CAFC appeal from any PTAB FWD; no CourtListener Federal Circuit docket exists for a PTAB appeal of this patent.
• Defensive value: N/A. The practical consequence is that you cannot free-ride on anyone else's invalidity work and you cannot point the court to a Board ruling that already killed a claim. If you want PTAB relief, you must file first.

Near-miss hits — explicitly not this patent (do not conflate)

Because a negative finding is only as good as the search hygiene behind it, here is what the searches actually surfaced, with the numbers reported exactly as they appeared:

None of these is US 10,900,770, and none of them is a Canon-owned imaging-plane-phase-difference patent. I did not auto-correct any number.

---

Strategic summary

Claim status of 10,900,770: the entire patent is UNTESTED. There are no canceled claims, no sustained claims, and no claims that survived a motion to amend (a contingency that cannot arise absent an IPR). Every claim the patent issued with on 2021-01-26 remains in force, including the device claim recited in the SUMMARY OF THE INVENTION — the distance measuring device with an optical filter having a first region of a first spectral transmittance characteristic and second/third regions of a second, longer-wavelength spectral transmittance characteristic, a first pixel portion with a first photoelectric conversion unit receiving light through the first region, a second pixel portion with second and third photoelectric conversion units receiving light through the second and third regions, and an acquiring unit producing distance information from the parallax of image data based on the second and third photoelectric conversion units' outputs. (Note: the authoritative excerpt provided to me is truncated before the printed claim set, so I am not enumerating dependent claim numbers — I will not guess at claims 2–N.)

Estoppel landscape: empty, and that cuts both ways. Because no petitioner has ever reached an FWD, § 315(e)(2) estoppel has never attached to anyone. For a defendant today, that means the full field of prior art is available — no ground is foreclosed as "raised or reasonably could have been raised." It also means you would be the first mover: if you file an IPR and it is instituted and reaches an FWD, § 315(e)(2) will bind you (and your privies/RPIs) as to every ground you raised or reasonably could have raised. In other words, a well-drafted single petition is worth more here than a scattershot one, because your first petition defines your own estoppel ceiling. Grounds not available in IPR (e.g., § 101 eligibility, best-mode, and § 112 written-description/enablement) remain district-court-only tools in any event.

Pattern signals: no pattern, because there is no activity. No petitioner has filed multiple IPRs at this patent. The patent owner (Canon Kabushiki Kaisha) has not had to defend a PTAB appeal because there has been no FWD to appeal. No defensive aggregator (Unified Patents, RPX, etc.) appears anywhere in the chain for this patent. Combined with the separate litigation finding of no district court suit and no CAFC appeal on 10,900,770, the picture is a patent that is simply not being worked — consistent with it being an internal Canon portfolio asset in the busy Canon imaging-plane-phase-difference family (cf. Nobayashi-family patents such as 10,690,495, 11,077,569, 11,808,607), rather than a monetization vehicle.

Calibrate expectations accordingly. The absence of PTAB activity is itself information, but read it correctly: it does not mean the claims are strong. Well-asserted patents eventually attract IPRs precisely because someone is being sued. A patent with no IPRs and no suits is usually a patent nobody has had a reason to attack. If that changes — if you or a client receives a demand letter — the asymmetry is real: you would face a completely untested claim set with no adverse claim-construction history.

---

Recommended next steps

If you are a defendant or demand-letter recipient:

1. Do not expect to cite a PTAB outcome — none exists. There is no FWD to quote, no canceled-claim disposition, and no CAFC mandate. Any argument that "the PTAB already invalidated claim X" would be fabricated; the record contains nothing of the sort.
2. Budget for a first-filed petition. With no prior petition, there is no General Plastic / § 314(a) follow-on risk against you and no Fintiv history to work around — but also no prior art-mapping precedent. The imaging-plane-phase-difference space is dense: JP 5917207 is cited on the face of the patent as admitted prior art (the background section expressly describes the imaging-plane phase-difference method of Japanese Patent No. 5917207), and the specification itself frames the invention as a spectral-transmittance-partition improvement over it. That citation is the natural starting point for a § 103 combination.
3. Preserve the doctrines IPR cannot reach. § 101 and § 112 challenges must be made in district court; if you intend to assert them, do not assume an IPR filing substitutes.
4. Watch the one-year bar. § 315(b) runs one year from service of a complaint alleging infringement — no clock has started because no complaint exists.

If you are the patent owner or a licensor:

• The patent is intact and untested. That is a clean slate for licensing narratives, but it also means there is no PTAB validation of the claims to cite. Consider that a prospective licensee or an accused infringer has an unrestricted prior-art field against you.

If you need to close the verification gap (recommended):

• I could not complete an authoritative full-text docket query in this session. To convert a negative finding into a defensible one, run a direct search for "10,900,770" on USPTO PTAB E2E (https://ptacts.uspto.gov/ptabweb/) and USPTO Patent Center (https://patentcenter.uspto.gov/) for the application US 16/353,573, plus CourtListener (https://www.courtlistener.com/) for any Federal Circuit appeal of a PTAB decision on this patent. Public databases can lag filings, and patent numbers occasionally appear only in petition exhibits — so treat this as a high-confidence but not airtight negative.

Caveats on completeness, stated plainly:

• This conclusion rests on the ODP structured block (which reports no proceedings) plus web searches that returned nothing on point. I did not fabricate any proceeding number, judge name, FWD disposition, or docket number, and I have not asserted any claim-level outcome because none exists.
• I have not verified the patent's total claim count because the authoritative text supplied to me truncates before the printed claims; I therefore make no statement about claims 2 through N.
• All dates above use YYYY-MM-DD where dates are given.

I'll research the assignment record for US 10,900,770, starting with the USPTO assignment databases.

Ownership & Assignment Analysis — US Patent No. 10,900,770 B2

Patent: "Distance measuring device, imaging apparatus, moving device, robot device, and recording medium"
Appl. No.: 16/353,573 (filed 2019-03-14) · Issued: 2021-01-26 · Priority: JP 2018-059938 (2018-03-27)
Searched: USPTO Assignment Center / Patent Assignment Search (by patent no. and by appl. no. 16/353,573), Google Patents legal events, PEDS/ODP, patents-family metadata (PatentGuru, Patent Leaderboard).

Data-availability note up front (important for how to read the timeline): I was able to confirm that an assignment was recorded, and its recording date and parties, but the reachable sources did not return the reel/frame number or the named correspondent for this specific record. Per the operating constraints I will not invent a reel/frame or an attorney name. Every reel/frame field below is marked NOT RETRIEVED rather than guessed. This limits the strength of Signal 3 (repeat correspondent) and is the reason the verdict is not stronger.

---

Inventors

Pattern notes:

• No unusual departure pattern. Nobayashi's inventorship record is a dense, continuing Canon portfolio: 43 issued US patents, all Canon-assigned, including US 12,315,179 (appl. 18/410,803, filed 2024-01-11) and US 12,498,214 (appl. 18/489,963, filed 2023-10-19) — i.e., he was still filing Canon-assigned applications five-plus years after the 2019 filing here. A "inventors leave within 12 months → fire-sale" precursor is therefore not present.
• Related Nobayashi/Canon filings name different outside counsel over time — Venable LLP (e.g., US 10,267,623) and Canon U.S.A., Inc. IP Division (e.g., US 12,315,179) — which is normal in-house/outside-counsel rotation, not an assignment-chain tell.

---

Original assignee

Canon Kabushiki Kaisha (Canon Inc.), Tokyo, Japan — named as assignee on the issued patent and as "Current Assignee" by Google Patents.

• Line of business: operating manufacturer of imaging, optical, and industrial equipment. The claims here (division-of-pupil optical filter + imaging-plane phase-difference photoelectric conversion units) sit squarely in Canon's core sensor/AF technology — the same imaging-plane phase-difference principle Canon commercializes in its Dual Pixel-family autofocus sensors, in on-vehicle cameras, and in industrial/robot vision products.
• Did they ship a product embodying the claims? Canon is an operating company that ships image sensors and camera/imaging systems practicing imaging-plane phase-difference AF; the specification itself targets vehicle-mounted (FIGS. 13A–13B) and robot-end-effector (FIG. 15) uses. I cannot point to a single retail SKU confirmed to practice these specific claim limitations — treat "ships embodying product" as highly likely but not claim-charted.
• Status: operating; publicly listed; no bankruptcy, receivership, or restructuring event identified. No 8-K/10-K disclosure of a patent sale involving this patent.

---

Assignment timeline

There is exactly one recorded assignment for this patent — the routine inventor-to-employer conveyance recorded pre-issuance. There is no post-issuance assignment, so the original assignee is, as far as the record shows, the current owner.

• Execution date: NOT RETRIEVED (conventional practice for this family is execution on/around the 2019-03-14 US filing date) / recorded 2019-06-06 — Reel NOT RETRIEVED / Frame NOT RETRIEVED
  • Conveyance: Assignment — "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)"
  • Assignor: Kazuya Nobayashi (individual inventor)
  • Assignee: Canon Kabushiki Kaisha (Canon Inc.)
  • Correspondent: NOT RETRIEVED for this record. For context only (not a finding for this reel): canon-family recordings in this inventor's portfolio have been filed through Venable LLP and Canon U.S.A., Inc. IP Division. Because I cannot tie either firm to this record, no recurrence signal is claimed.
  • Context: intra-corporate — inventor's statutory assignment to his employer; the standard step that perfects title before grant. Not a fire-sale, not a securitization, not a transfer-to-asserter.

Other recorded legal events (not assignments, listed to show the record is otherwise empty of transfers):

Contradiction check with the earlier litigation section: none. That section concluded no litigation naming US 10,900,770, and separately flagged near-miss numbers (10,790,703 Koji IP v. Renesas; 10,659,885 / 10,820,117 Taction v. Apple; Canon toner-cartridge and Lodsys/Canon suits). Nothing in the assignment record contradicts that finding — and the absence of any post-issuance assignment is consistent with there being no transfer-to-asserter precondition for a suit.

---

Timeline diagram

timeline
    title Ownership of US 10900770
    2018 : JP priority application filed by Canon
    2019 : US application filed by Canon
         : Inventor assigns rights to Canon Kabushiki Kaisha
         : Pre-grant publication
    2021 : Patent granted to Canon
    2039 : Anticipated expiration

---

NPE / troll-pattern signals

---

Verdict

Insufficient data (only the original assignment is recorded).

Reasoning: the entire recorded chain is a single inventor-to-employer assignment, recorded 2019-06-06 (reel/frame not retrieved) from Kazuya Nobayashi to Canon Kabushiki Kaisha, with no subsequent transfer, no security interest, no merger, and no change of name (Google Patents legal events for US 10,900,770; consistent with the earlier finding of zero litigation). Because the only assignee is a large operating manufacturer, every NPE-pattern signal is either "not present" or, for the correspondent test, unresolvable — so the rubric's Insufficient data bucket is the honest fit rather than "Operating-company assertion," which would require an actual assertion (there is none) or a claim-charted embodying product (I could not verify one). The one open item that could theoretically upgrade confidence in either direction — the correspondent of record on the 2019-06-06 recording — should be pulled directly from the record document.

Verification links:

• USPTO Assignment Center (search: 10900770, or 16/353,573): https://assignmentcenter.uspto.gov/
• Legacy Patent Assignment Search (same underlying records): https://assignment.uspto.gov/patent/index.html#/patent/search?q=10900770
• Google Patents legal events for this patent: https://patents.google.com/patent/US10900770/en

To close the open field: request the assignment abstract of title / the recorded assignment document for appl. 16/353,573 to capture the reel/frame and the correspondent attorney of record; I could not obtain either from the sources reachable in this session and have deliberately left those cells blank rather than infer them.

I'll search for the USPTO record of US 10,900,770 and its cited prior art.

Let me pull the citation listing directly from the patent record pages.

Prior-Art Analysis — U.S. Patent No. 10,900,770 B2

Note on sources used: The authoritative full text you supplied is treated as the primary source. I also pulled the front-page "References Cited" listing reproduced on the patent's public record (Justia, https://patents.justia.com/patent/10900770), cross-checked against Google Patents (https://patents.google.com/patent/US10900770/en). My tool session was truncated before I could re-download the USPTO Patent Center "References Cited" tab directly, so the list below is reported as it appears on the patent face, but should be re-verified against USPTO Patent Center for application 16/353,573 (see Caveats at the end).

---

1. Identity Verification (exact number, no auto-correction)

The number was searched literally as 10900770. Results for similar numbers (e.g., 10,790,703; 11,000,207; 10,906,285; 10,938,012) appeared in the search stream and are not this patent; they are excluded.

---

2. Cited References on the Face of US 10,900,770

The front page ("References Cited / Referenced Cited") lists the following. No examiner-cited foreign publication other than JP 5917207 was surfaced, and no "Other Publications" (NPL) entry was surfaced.

U.S. Patent Documents

Foreign Patent Documents

Also cited inside the specification (Description of the Related Art): "Japanese Patent No. 5917207 discloses an imaging apparatus using an imaging plane phase difference method, which can acquire not only an image to be used for the recognition process but also a distance." This is the same JP 5917207 B2 above — cited both in the body and (per the front page) in the reference list.

---

3. Claims at Issue (for the § 102 mapping)

Independent and key dependent claims, read from the granted text:

• Claim 1 (independent) — Distance measuring device comprising: (a) an optical filter having a first region with a first spectral transmittance characteristic and second and third regions with a second spectral transmittance characteristic transmitting longer-wavelength light than the first; (b) a first photoelectric conversion unit in a first pixel portion receiving light through the first region; (c) second and third photoelectric conversion units in a second pixel portion receiving light through the second and third regions; (d) at least one processor + memory programmed as an acquiring unit that acquires distance information corresponding to the parallax of image data from the second/third unit outputs.
• Claim 2 (independent) — a related device claim (structure of filter + first photoelectric conversion unit), recited in the Justia rendering as a second "comprising" claim.
• Claim 3 (dep. claim 2) — first region transmits λ1–λ2; second/third regions transmit λ3–λ4; λ1 < λ3 and λ2 < λ4.
• Claim 4 (dep. claim 2) — λ2 < λ3 (i.e., the ranging band lies entirely above the visible band).
• Claim 5 (dep. claim 3) — λ3 ≥ 450 nm or λ4 ≤ 1100 nm.

The claim set therefore has two independent focal points: the split-optical-filter / split-pupil arrangement and the spectral-partition (color-separation) limitation keyed to the imaging-plane phase-difference ranging signal.

---

4. Per-Reference Analysis and Potential § 102 Mapping

Legal frame: The March 27, 2018 priority date is post-AIA, so 35 U.S.C. § 102(a)(1) (patented/published before the effective filing date) and § 102(a)(2) (earlier-effective-filing-date U.S. patent/application publication) govern. Note references are cited by the examiner as art; citation alone does not establish anticipation — it establishes that the examiner considered them relevant.

4.1 US 9,456,119 B2 — Hamano — issued Sept. 27, 2016

• § 102 basis: § 102(a)(1) — patented more than 1 year before the 2018 priority date. Statutorily qualifies as prior art.
• Brief description: A granted U.S. patent in the imaging/imaging-sensor field (cited by surname "Hamano" on the face). Its presence in the citation list of an imaging-plane-phase-difference patent indicates it was relied on as background art for the sensor/pupil-division or color-filter architecture.
• Potential anticipation of claims: Most plausible against claim 1 / claim 2 if — and only if — it discloses a pupil-splitting optical element with at least two distinct spectral-transmittance regions co-located with two (or more) photoelectric conversion units in a single pixel. That is the classic imaging-plane phase-difference architecture. It would not anticipate the spectral-partition limitations of claims 3–5 (the λ1<λ3<λ2<λ4 / λ2<λ3 / 450–1100 nm requirements) unless it expressly teaches the long-wavelength-shifted ranging sub-band.
• Assessment: Potentially anticipatory as to claim 1/claim 2 structure only; my confidence is low-to-moderate and I could not verify its technical content within this session.

4.2 US 2018/0010903 A1 — Takao — published Jan. 11, 2018

• § 102 basis: § 102(a)(1) — published before the March 27, 2018 effective filing date. Statutorily qualifies as prior art as of its publication.
• Brief description: A U.S. published application in the imaging/sensor field (cited by surname "Takao"), cited as background art in the same sensor/pupil architecture space.
• Potential anticipation of claims: Same logic — could be mapped against claim 1 / claim 2 if it discloses the split-optical-filter with a shorter-wavelength imaging region and longer-wavelength ranging regions feeding separate photoelectric conversion units. If it is a color-filter / pixel array reference only (without the phase-difference parallax acquisition), it would at most support a § 103 combination, not anticipation.
• Assessment: Potentially anticipatory only for the structural independent claims; § 103 combination is the more likely examiner posture.

4.3 US 2018/0136477 A1 — Moriuchi — published May 17, 2018

• § 102 basis: Important timing point. Its publication (May 17, 2018) is after the March 27, 2018 priority date, so it is not § 102(a)(1) art. It can only operate under § 102(a)(2) as a U.S. application publication having an earlier effective filing date (typical 18-month publication implies a filing around late 2016), and it must not fall within a § 102(b)(2)(C) common-ownership exception (relevant because it is likely Canon-assigned, like the patent at issue).
• Brief description: A U.S. published application (cited by surname "Moriuchi") in the imaging field, cited as background art.
• Potential anticipation of claims: Because it is expressly identified on the face of the patent, the examiner evidently considered it material to at least the structure of claim 1 / claim 2 (pupil-division pixel architecture). The spectral-transmittance limitations of claims 3–5 would only be anticipated if Moriuchi expressly discloses a ranging sub-band offset to longer wavelengths.
• Assessment: Potentially anticipatory as to the structural claims under § 102(a)(2); verify its filing date and any § 102(b)(2)(C) common-ownership issue.

4.4 JP 5917207 B2 — registered/published May 2016

• § 102 basis: § 102(a)(1) — published (JP registration/publication) before the priority date. (Note: JP 5917207 is a Japanese patent number, not a U.S. number; the specification cites it by number verbatim.)
• Brief description (from the patent's own text): An imaging apparatus using an imaging-plane phase-difference method, which can acquire not only an image for a recognition process but also a distance. This is the closest § 102 reference cited in the specification background.
• Potential anticipation of claims: This is the most substantive § 102 candidate against claims 1 and 2, because it appears to disclose the core method — split pupil regions, two image signals, parallax → defocus → distance. However, the novelty of the patent-at-issue is expressly framed against JP 5917207: the distinguishing feature is the three-region optical filter where the first (imaging) region has a visible-band transmittance and the second/third (ranging) regions have a longer-wavelength-shifted transmittance. Unless JP 5917207 discloses that spectral partition, it does not anticipate claims 1–2 and certainly not claims 3–5.
• Assessment: Highly relevant, but on the specification's own characterization it is distinguished by the wavelength-partition feature; most likely a § 103 or background-art citation rather than an anticipation reference.

---

5. Bottom Line

The most relevant prior art among the cited references is JP 5917207 B2 (imaging-plane phase-difference imaging apparatus; the reference the specification itself distinguishes), followed by US 9,456,119 B2 (Hamano) and US 2018/0010903 A1 (Takao) as structural background art. No cited reference appears, on the information available, to anticipate the wavelength-partition limitations of claims 3–5, which read as the genuine point of novelty (ranging band shifted long relative to the imaging band: λ1<λ3<λ2<λ4; λ3 ≥ 450 nm; λ4 ≤ 1100 nm).

---

6. Caveats and Explicit Flags

1. List may be non-exhaustive. My session was truncated before I could open the USPTO Patent Center "References Cited" tab for application 16/353,573. The list above is taken from the patent's reproduced front page; the full face may include additional cited U.S. documents, foreign documents, or NPL entries I did not capture. Recommend: confirm via USPTO Patent Center (https://patentcenter.uspto.gov/, app. 16/353,573) and Google Patents "Patent Citations."
2. I could not verify the technical content (titles, assignees, disclosed subject matter) of US 9,456,119 B2, US 2018/0010903 A1, and US 2018/0136477 A1 in this session. My § 102 mappings for those three are inferential and explicitly flagged as low-to-moderate confidence — do not treat them as claim-chart conclusions.
3. No auto-correction applied. All identifiers are reported exactly as found: "10900770," "US 2018/0010903 A1," "US 2018/0136477 A1," "US 9,456,119 B2," "JP 5917207." Near-miss numbers appearing in search results (10,790,703; 11,000,207; 10,906,285; 10,938,012) were deliberately not substituted.
4. Date contradiction to flag (per your instruction): The previously generated Litigation Summary lists a Taction Technology v. Apple verdict dated September 25, 2026, yet the task header states "Current Date: April 26, 2026" (and the system date is September 30, 2026). A September 25, 2026 event cannot precede an April 26, 2026 "current date." This does not affect the 10,900,770 analysis (that case concerns unrelated patents), but it is a genuine internal inconsistency in the prior sections.
5. No litigation/PTAB activity on 10,900,770 was found (consistent with the earlier Litigation Summary) — so there is no PTAB or court claim construction or invalidity ruling that would give a petitioner's or court's own prior-art mapping for the claims.

If you want, I can (a) re-query specifically for the full USPTO "References Cited" face of 16/353,573 to close the exhaustiveness gap in item 1, and (b) run targeted lookups on the three U.S. references to convert the provisional § 102 mappings into verified claim-by-claim comparisons.

I'll retrieve the prior-art-of-record and related references for this patent before building the § 103 analysis.

Let me pin down the two references cited of record and look for other family/prior-art listings.

Obviousness Analysis — U.S. Patent No. 10,900,770 B2

Distance measuring device, imaging apparatus, moving device, robot device, and recording medium (Nobayashi; Canon K.K.)

Scope note before we begin. You asked me to use the Prior Art section of the page. I want to be transparent about what that section actually yielded. The Google Patents page includes only two prior-art citations of record (US 2018/0010903 A1 (Fujifilm) and US 2018/0136477 A1 (Toshiba)) per the earlier-generated bibliographic section, and the specification itself cites JP 5917207 as admitted prior art. My searches did not return full-text copies of the two US citations of record, so I analyze them at the level of what their numbers, assignees, dates, and field establish, and I flag where verification is required rather than inventing disclosures. Everything below is analysis, not a legal opinion; obviousness is ultimately a legal conclusion.

---

1. Legal framework applied

• Governing statute: AIA 35 U.S.C. § 103 (effective filing date 2018-03-27; no pre-AIA issues).
• Controlling case law: KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007); Graham v. John Deere Co., 383 U.S. 1 (1966); In re Keller, 642 F.2d 413 (CCPA 1981); In re Aller, 220 F.2d 454 (CCPA 1955) (routine optimization of a known parameter/range); In re Boesch, 617 F.2d 272 (CCPA 1980) (recited results must be shown to be unexpectedly superior).
• Rationales available under MPEP 2143 / KSR: (A) combine prior art elements according to known methods to yield predictable results; (B) simple substitution of one known element for another; (C) use of known technique to improve similar devices in the same way; (D) applying a known technique to a known device ready for improvement; (E) "obvious to try" where a finite number of identified, predictable solutions exist; (F) design incentives/market forces; (G) one of the enumerated TSM rationales.
• Key doctrinal lens here: this is a classic "arrangement of old elements" case. If the color-image channel and the ranging channel operate independently and the asserted benefits are additive rather than synergistic, the combination is highly vulnerable under KSR's "predictable results" line.

---

2. Claim 1 — element-by-element breakdown

Claim 1 (as reproduced verbatim in the patent's own Summary and as drafted in the earlier section) breaks into five elements:

Observation on claim breadth: Claim 1 does not recite (a) that the first region is central or flanked; (b) any numeric wavelengths; (c) any color-filter array; (d) any area ratio; (e) that the second and third regions are pupils used for phase-difference ranging specifically. It is a bare two-band pupil-allocation claim plus generic processor language. That breadth is important — the narrower, arguably inventive details (λ₃ ≥ 450 nm, λ₄ ≤ 1100 nm, area ratios, the movable diaphragm 601, the pattern projector) live, if anywhere, in dependent claims.

---

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

A POSITA at the 2018 priority date would hold a B.S. in electrical engineering, optics, or applied physics (or equivalent) with 2–4 years of experience in CMOS image-sensor design and/or camera optics, or an M.S. with less. That person would be familiar with: Bayer CFAs; on-chip microlenses; pupil-division phase-difference AF; lateral/axial chromatic aberration of glass lens systems; Si photodiode spectral response (usable to ~1000–1100 nm, absorption coefficient falling with wavelength); interference/dielectric multilayer filters; and normalized cross-correlation / SSD block matching for disparity estimation. This is a modest skill level — which cuts against patentability.

---

4. Prior art inventory

4.1 Prior art of record / admitted prior art (highest confidence)

Rigor flag: I could not retrieve the full text of US 2018/0010903 A1 or US 2018/0136477 A1 in this session. Do not treat the "what it establishes" column for those two as verified disclosures — those are inferences from number, assignee, and field. Any formal obviousness contention would need their actual paragraphs cited.

Anti-conflation flag (important): Several of my searches returned other search reports and IPR papers containing prior-art lists — e.g., CN 102763232 A / US 2012/043569 A1 / US 2012/043573 A1 (Mitsuishi, Toshiba), which appear in a search report for PCT/CN2014/080435, and Janssen '708/Kubota from unrelated PTAB petitions. These are not references of record in '770 and I am not treating them as such. I note them only as evidence of the crowdedness of the field.

4.2 Candidate prior art located but not confirmed on the face of '770 (medium confidence, must be verified)

4.3 What the specification admits is known (these admissions shrink the gap)

Courts and examiners can use applicant admissions. From the '770 text:

• "the defocus amount can be converted into a defocus amount by using a predetermined conversion coefficient … converted into information about distance … by using an image-forming relation" → defocus-to-distance is routine.
• "a method similar to the parallax amount detection method in a stereo imaging apparatus"; "SSD (Sum of Squared Difference)" → disparity estimation is routine.
• "This correction method is well known" (light-amount balance between the pair).
• "a light shot noise generated by the photodiode is dominant" → the noise problem is recognized prior-art knowledge.
• "Many glass materials that are transparent in the visible wavelength band have a normal dispersion in which the refractive index increases as the wavelength becomes shorter … as the wavelength becomes shorter, the chromatic aberration of the imaging optical system 102 becomes longer" → the chromatic-aberration physics is textbook.
• "A photodiode … is configured by, for example, Si" → silicon response is baseline.
• "a well-known technique is used for setting or correcting the spectral sensitivity" → spectral engineering is well known.

---

5. Ground 1 — Claim 1 obvious over JP 5917207 in view of a band-selective filter reference (e.g., Toshiba US 2013/0242087 / US 9,479,760) and in further view of a pupil-division color-filter reference (e.g., US 2015/0109498 or US 9,172,925)

5.1 Where each reference lands on the claim

The single unmet limitation is 1.4's relative spectral relationship (ranging sub-pupils pass longer wavelengths than the imaging sub-pupil). Everything else is squarely in the art. That means the obviousness question reduces to a narrow one: would a POSITA have been motivated to allocate the long-wavelength portion of the spectrum to the ranging sub-pupils and the visible portion to the imaging sub-pupil, with a reasonable expectation of success? The answer, on the record, is yes — for at least six independent, mutually reinforcing reasons.

5.2 Articulated motivations to combine (the heart of the § 103 case)

(M1) Same field, same problem, same solution space. All references are in solid-state imaging devices / distance measurement through a shared taking lens. Combination of references "directed to the same problem" is the paradigm case. In re Keller; MPEP 2143.01.

(M2) Both references teach the same two error mechanisms the patent addresses — and both are admitted known.

• Shot noise: the spec concedes shot noise "is dominant," and that broadening the received band is the cure ("an amount of light received by the distance measuring pixel portion can be increased by receiving light in the wavelength band as wide as possible"). A POSITA seeking to reduce shot-noise-induced disparity error would broaden the ranging band as a matter of course.
• Axial chromatic aberration: the spec concedes the dispersion physics and the fact that CA error is larger at short wavelengths. A POSITA seeking to reduce wavelength-dependent focus offset in a phase-difference distance measurement would exclude short wavelengths from the ranging channel as a matter of course.

These two design moves together produce exactly the claimed λ₃–λ₄ band: λ₃ pushed above the blue end (≈450–490 nm) and λ₄ extended into the red/NIR. Element 1.4 is thus the predictable output of two independently known optimizations, not an inventive insight. This is the strongest single point in the analysis.

(M3) Physics of silicon supplies the "why longer." A POSITA knows Si photodiodes have high quantum efficiency through the red/NIR and cut off near 1100 nm (which is precisely why '770's own dependent disclosure caps λ₄ at ≤1100 nm, preferably <1000 nm, citing "the absorption coefficient of Si is smaller toward the longer wavelength side and the crosstalk between adjacent pixels becomes larger"). Red-shifting a ranging channel to harvest more carriers across a wider band is a predictable, well-understood engineering choice — and the patent's own λ₄ rationale is a record-based admission of that knowledge.

(M4) The art itself already teaches putting a band-selective filter in the ranging instrument's optical path. US 9,479,760 (Toshiba) interposes filter 104 ahead of the microlens array of a ranging imaging device expressly "to block unnecessary light." Once a POSITA has (i) a filter in the ranging light path and (ii) separately known that short-wavelength light should be excluded and long-wavelength light harvested, placing the filter's passband differently in different pupil regions is a mere substitution/arrangement of known elements doing what they were designed to do (MPEP 2144.04; KSR). The result — a pupil region that passes a longer band and a pupil region that passes a shorter band — is entirely predictable.

(M5) The color-reproduction requirement requires the two channels to be spectrally separated, which supplies the motivation. The spec's own statement of the prior-art problem: if the imaging pixel receives NIR, "each pixel having spectral sensitivity of red, green, and blue has a sensitivity also in the near infrared wavelength range," so "an object … having a high spectral reflectance in the near infrared wavelength range may be recognized as white even though it is black." A POSITA who wants (a) more light for ranging and (b) clean color, and who already has spatially distinct pupil regions, is led directly and predictably to assign the long band to the ranging pupils and the visible band to the imaging pupil. Far from being inventive, the claimed arrangement resolves a known conflict in the art — a hallmark of obviousness.

(M6) "Obvious to try" with a finite, predictable solution set. The design space for reducing (noise + CA) error in a pupil-division range channel is small and enumerable: (i) widen the band; (ii) shift the band redward; (iii) set the band edges by balancing light budget against CA error; (iv) choose where in the optical train to put the spectral split. The patent does no more than select within that finite set, and its own disclosure describes the selection as a balance: "it is necessary to set the area ratio of the three regions … based on the distance measuring range and the minimum object illuminance" and "the wavelength λ₃ can be determined taking into consideration the chromatic aberration … and the distance measurement accuracy required." Where the reference discloses a recognized trade-off, selecting a working point is routine optimization. In re Aller; In re Boesch.

5.3 No synergy — the KSR "predictable result" problem

This deserves its own header because it is likely dispositive.

The two claimed benefits are produced by two physically independent channels:

• Color reproduction comes exclusively from the first PCU 211 seeing the visible band (spec: image processing "performed on image signals output from the imaging pixel portions 103a"; NIR exclusion protects color).
• Ranging accuracy comes exclusively from the second and third PCUs 212/213 seeing the longer band (spec: distance information "generated based on the viewpoint image data by using the output signals of the photoelectric conversion units 212 and 213 without using the output signal of the photoelectric conversion unit 211").

There is no claim element in which one channel benefits from the other's spectral choice. The '770 specification itself effectively concedes independent action: in the FIG. 3E alternative, the imaging pixel's other two conversion units are simply not read ("the imaging element 103 is wired so as not to output each of the signals of the photoelectric conversion unit 212 and the photoelectric conversion unit 213"), and vice-versa. Under KSR, a combination that does no more than yield the predictable, additive results of its known parts is obvious. Absent evidence that the specific spectral partition yields a result greater than the sum of "visible imaging band" + "redward ranging band," the synergy argument fails.

---

6. Ground 2 — Alternative primary references (fallback grounds)

If the examiner/defendant does not want to rely on a Japanese granted patent as the primary, the same architecture is available in English-language US publications:

• Ground 2a: US 2015/0109498 A1 (Fujifilm) as primary — imaging element with first/second phase-difference pixels receiving pupil-divided light from first/second areas plus color filters and a phase-difference detection device — in view of US 9,479,760 / US 2013/0242087 (Toshiba) for the band-selective filter in the ranging optical path, and in view of the admitted optics (Si response, dispersion).
• Ground 2b: US 9,172,925 B2 (Fujifilm) as primary — Bayer array with phase-difference pixel cells 51R/51L receiving light through different pupil areas — in view of US 6,714,732 for the light-amount-compensated correlation-based disparity computation (element 1.3), plus the same band-filter and optics combination.
• Ground 2c: US 2013/0242087 A1 (Toshiba) as primary — an expressly named "Distance Measuring Device" with a solid-state imaging device and compensated multi-pixel sensing — in view of a pupil-division color-filter reference.

Each of these grounds keeps the same logical core: pupil-division ranging (known) + band-selective filtering in the optical path (known) + two known error mechanisms with known directional cures (admitted) → the claimed spectral allocation (predictable).

---

7. Dependent claims — obviousness assessment

I do not have the verified granted dependent claim set (flagged in the earlier section), so I analyze the limitations that appear in the specification as likely claim fodder.

---

8. Claims 13–15 (moving device, robot device, storage medium)

• Claim 13 (moving device = distance measuring device + control unit controlling a drive unit using the image and distance information). Obvious. Vehicle-mounted ranging for collision avoidance is an old and crowded field (the '770 classification itself lists G01S17/931, "Lidar systems … for anti‑collision purposes of land vehicles," and B25J19/021). Applying the claimed sensor arrangement to a vehicle is using a known technique to improve a similar device in the same way (MPEP 2143.01(III); KSR). The patent's own FIGS. 13A–13B describe nothing beyond conventional ADAS logic (recognize preceding vehicle, compare distance to threshold, brake or alarm). No new structural or functional relationship between the sensor and the vehicle is claimed. Strongly obvious.
• Claim 14 (robot device = distance measuring device + robot arm + robot hand + control unit computing position/orientation of a workpiece from image + distance). Obvious. Robot vision with combined 2D image + depth for bin-picking/grasping pose estimation is a decades-old field, and FIG. 15 of '770 is a textbook articulated-arm pick-and-place. Field-of-use/obvious-design-choice. Strongly obvious.
• Claim 15 (non-transitory storage medium storing a program making a computer function as the acquiring unit). Obvious. Purely a program-product claim covering the same logic implemented on a generic processor. Once the device claim falls, the medium claim falls with it. (Note: the claim as described requires "non-transitory," so In re Nuijten's signal-claim problem is avoided, but that does not import patentability — Beauregard-style claims are assessed on the merits of the underlying method/apparatus.)
• "Imaging apparatus" (in the title, and the JP counterpart's first/second color filter recitation). Per the earlier section, it is unconfirmed whether a separate independent imaging-apparatus claim exists in the granted US set. If it recites only "first and second color filters" on the two pixel types with the spectral relationship, it is more obvious, not less — the color-filter-on-pixel version is the ordinary, art-known way to implement a spectral split, and the patentee itself concedes that implementing the split at the pixel (via dielectric multilayer color filters, FIGS. 11–12) is an alternative configuration.

---

9. Anticipated patentee rebuttals, and why they are weak

---

10. Overall conclusion and confidence

Conclusion (medium-high confidence): Claim 1 of US 10,900,770 is likely obvious under § 103 over the admitted imaging-plane phase-difference prior art (JP 5917207) in view of a band-selective optical filter in a ranging imaging device (Toshiba US 2013/0242087 / US 9,479,760 / WO 2014/054290) and a pupil-division color-filter phase-difference reference (Fujifilm US 2015/0109498 / US 9,172,925 / US 8,723,992), further in view of the admitted knowledge of Si spectral response and lens dispersion. The claimed spectral allocation is the predictable output of two independently known error-correcting design moves (broaden the ranging band for shot noise; exclude short wavelengths for chromatic aberration), performed on an architecture the patentee itself concedes was already known. Element 1.4 is the only limiting feature not squarely met, and the motivation to supply it is strong and multiply grounded.

Claims 13, 14, and 15 are even more plainly obvious — field-of-use/design-choice applications of a known sensor, and a program-product claim reciting generic hardware.

The strongest patentability hook left to the patentee is the numeric λ₃ working points (≥450 nm; 470 nm; 490 nm) and, secondarily, the λ₄ ≤ 1100/<1000 nm limit — and even those are routine-optimization candidates because the specification supplies the reason for each number from known physics. A robust invalidity position should therefore target these ranges with art that discloses or suggests a red-shifted phase-difference/NIR ranging channel, or with art showing the trade-off curve.

---

11. Explicit limits on this analysis (read before relying on it)

1. I could not retrieve the full text of the two references cited of record (US 2018/0010903 A1; US 2018/0136477 A1). My characterization of them is inferential. Pinpoint citations to their paragraphs must be obtained before any ground is asserted.
2. I could not verify the complete granted claim set — total claim count, dependency structure, and whether any claim was amended during prosecution. Dependent-claim analysis in § 7 is based on the specification's disclosure, which is a proxy, not the granted claims.
3. Numbers were not auto-corrected. The references in §§ 4.1–4.2 are reported exactly as found. In particular, do not conflate 10,900,770 with neighbors surfaced in searches (10,790,703; 10,659,885; 10,820,117; 10,891,608-style numbers), nor with prior-art lists lifted from other search reports (e.g., the CN 102763232 / US 2012/043569 / US 2012/043573 citations belong to PCT/CN2014/080435, and Janssen '708 / Kubota belong to unrelated PTAB petitions).
4. No litigation/PTAB/CAFC record for this patent was found (consistent with the earlier section), so there is no prior judicial claim construction or validity ruling to leverage.
5. The JP 7030592 B2 family divergence noted earlier (a machine translation suggesting λ₂ < λ₃ under the JP claim versus λ₁ < λ₃ < λ₂ < λ₄ in the US specification) remains unresolved and could matter if the family is used for claim-scope or priority purposes. Do not rely on it without reading the JP claims in Japanese.