US 4723149

I'll verify the record against live sources before summarizing.

US 4,723,149 — Verification Summary

Searches run: general web (Google Patents, FreePatentsOnline, Unified Patents, uspto.report, Espacenet, patentimages PDF), plus targeted queries for CAFC 2026 dockets citing 4,723,149. No CAFC or district-court 2026 activity surfaced. Note that a "no results" web search is not equivalent to a verified absence in a docket database — I do not have direct authenticated access to USPTO PatentCenter or the CAFC docket system, so treat the litigation/enforcement statement below as a search observation, not a certified clearance.

Bibliographic data

Discrepancy flagged (not auto-corrected): Unified Patents' portal (https://portal.unifiedpatents.com/patents/patent/US-4723149-A) lists priority 1985-05-07, application date 1986-05-04, and grant date 1988-02-01. Google Patents, the granted-PDF front page ("[45] Date of Patent: Feb. 2, 1988"), and Espacenet all give 1985-05-08 / 1986-05-05 / 1988-02-02. The off-by-one values on the Unified record appear to be portal artifacts; the granted document controls.

Abstract (as printed)

"An image forming apparatus for checking the copying of secret documents. The apparatus includes an original table on which original documents are placed. If the original document is a secret document, it is provided with a predetermined identification section. Scanning and image forming apparatus is provided which forms an image of the document when authorized to do so. In operation, secret document detecting apparatus detects whether or not a secret document identification section is present on a given document to be copied. First controlling apparatus is provided for driving the scanning apparatus prior to an image forming scan. Second controlling apparatus is provided for either enabling or disabling the image forming apparatus depending on whether a secret document is present. Third and fourth controlling apparatus is provided for driving the image forming apparatus in accordance with a displacement command of a light transmitting means which performs erasure of specified portions of the secret document by transmitting light through the original at specified erasure locations."

Independent claim 1 — plain language

There is exactly one independent claim (claim 1); claims 2–5 depend from it. Claim 1 is a single apparatus claim covering a copier that performs two combined functions: (a) detecting a secret document and blocking its copy, and (b) user-selected image-area erasure. Its elements:

1. Original table (platen) holding the original, which — if a secret document — carries a predetermined identification section.
2. Original scanning means movable along the table to optically read the original (the lamp/mirror carriage of the first embodiment).
3. Image forming means that puts the scanned image onto an image-forming medium (photosensitive drum → paper).
4. Secret-document detecting means mounted integrally with the scanning means, having a sensor that detects the identification section. In the disclosure this is a magnetic sensor (131) on the carriage, sharing the same positioning hardware as the trimming spot-light.
5. First controlling means that drives the scanning means before the copy operation and causes the sensor to two-dimensionally scan the platen (the carriage sweeps in y, the sensor is indexed in x on its own belt/stepping motor). Detection thus occurs irrespective of where the document sits on the glass.
6. Second controlling means that enables or disables the image forming operation based on the detection result, thereby preventing copying of the secret document (in the embodiment: halt the carriage, sound alarm 139, or transmit a remote alarm).
7. Light-transmitting means, integral with the scanning means, projecting light through the original (the spot light source 171) so the user can see the spot through the sheet.
8. Erasure-portion specifying means letting the user displace the transmitted-light spot and thereby designate an unnecessary portion of the original as the erasure portion, prior to image formation (keys 30a–30g; two designated points define a rectangle, or its inverse).
9. Third controlling means driving the scanning means in accordance with that displacement command to move the light-transmitting means to the commanded position (spot coordinates resolved from stepping-motor pulse counts).
10. Erasure-portion storage means storing the erasure position data (RAM 140 bit-map: "1" = erase, "0" = keep).
11. Image erasing means for selectively erasing image data transferred by the image forming means (the LED erasure array 150 placed between charger 11 and exposure region Ph, or alternatively between exposure Ph and developer 12).
12. Fourth controlling means that reads the stored erasure data out during the image-forming operation and supplies it to the image erasing means (rank-by-rank shift/ store register 161/162 and switch array 164, latched within θ₁/ω of drum rotation).

Practical reading of the claim: the inventive point asserted is shared hardware — one two-axis positioning system on the scanning carriage serves both the security sensor and the trimming spot-lighter — so copy protection and area-selection/erasure are obtained in a simple (non-duplicative) architecture.

Dependent claims

• Claim 2: the identification section is arranged along one side of the secret document.
• Claim 3: the identification section is arranged at a given position on the secret document (mid-document marking; requires finer sensor movement).
• Claim 4: the identification section is a magnetic medium and the sensor is a magnetic sensor (e.g., concealed magnetic tape).
• Claim 5: the first controlling means starts the detection scan from the position farthest from the original placement reference position on the platen (carriage returns from its rest position toward fixed scale 2₁), which the spec credits with high detection speed.

Context notes

• Cited prior art of record: US 3,977,785 (Xerox, "Method and apparatus for inhibiting the operation of a copying machine"), US 4,586,811 (Ricoh, "Confidential document reproduction prevention method"), US 4,486,090 (3M, "Copy machine with unauthorized copy prevention"), plus IBM Tech. Dis. Bull., "Preventing Copying of Classified Information," vol. 19, no. 4, Sep. 1976, p. 1469.
• The patent has been cited by 107 later documents (notably Canon's 1988-era anti-forgery family and Digimarc watermarking patents), which is why it is often described as an early ancestor of document-security/anti-copy detection in imaging devices.
• The three disclosed embodiments: (1) full combination of secret-document detection + trimming/erasure; (2) detection-only (trimming hardware and alarm optional); (3) improved dual-diameter spot light source with auxiliary surrounding light for easier visual confirmation.

Uncertainty statement: the bibliographic data, abstract, and claim text above come from the granted patent document rendered at https://patents.google.com/patent/US4723149/en and the granted PDF, cross-checked against FreePatentsOnline, uspto.report, and Espacenet. I found no CAFC 2026 appellate matter involving this number, but I could not query an authoritative case-management system; the patent lapsed in 1996 well before any 2026 docket, so a live 2026 appeal on this number would be anomalous and should be verified in PACER/CM-ECF or the CAFC docket if it matters to you.

I'll research litigation involving US patent 4723149 across litigation databases.

Let me search more specific litigation databases.

Let me try additional targeted searches.

Summary of Findings: Litigation Involving U.S. Patent No. 4,723,149

No litigation involving U.S. Patent No. 4,723,149 was identified in the sources searched.

I searched litigation-oriented resources (Unified Patents Portal, RPX Empower / litigation docket aggregators, CourtListener, and general docket sources) using the exact number 4,723,149 and variants ("4723149", "4,723,149"). No case — district court, ITC, or Federal Circuit — naming this patent as an asserted patent was returned.

What the searches did confirm (context, not litigation of the '149 patent)

Notably, the Google Patents page for the '149 patent shows only third-party citations — 3 cited references (US 3,977,785 / Xerox; US 4,586,811 / Ricoh; US 4,486,090 / 3M) and 45+ documents citing it (largely Canon and Digimarc anti-forgery/watermark patents). Citations are prior-art relationships, not litigation.

Important caveats

• No result should be treated as a negative finding of no litigation ever. Litigation from the patent's active period (roughly 1988–1996) predates most freely searchable electronic dockets. PACER coverage for district courts is generally complete only from the mid-to-late 1990s onward, and Unified Patents/RPX dockets skew toward post-2000 (especially post-2010) NPE activity. A case filed and resolved in, e.g., 1991–1994 would likely not appear in the sources searched.
• The Toshiba-related cases the search surfaced (e.g., Wang Laboratories v. Toshiba, Nos. 4,656,605/4,727,513; Monument Peak Ventures v. Toshiba, 8:19-cv-02181; Encryptawave v. Toshiba, 4:25-cv-00171; Lexar v. Toshiba) involve different patents entirely and are not relevant to the '149 patent. I excluded them rather than misattribute them.
• Given the patent lapsed for non-payment of maintenance fees in the 1990s and expired in 2006, it was never a long-lived or heavily asserted asset, which is consistent with the absence of any recoverable litigation record.

Bottom line: On the record available, there is no known litigation involving U.S. Patent No. 4,723,149. If you need a definitive legal-clearance-grade answer for the 1988–1996 active window, I would recommend a paid Westlaw/Lexis or Docket Navigator historical search, since those databases index older district court and ITC complaints that free sources omit. I do not want to represent the absence of search results here as proof that no suit was ever filed.

I'll verify the PTAB posture against live sources before writing anything up.

Let me run a second, differently-targeted set of searches to make sure the absence isn't a search artifact.

PTAB / AIA Trial Review — US 4,723,149

Source of the proceedings list: the structured "PTAB proceedings on file" block (USPTO Open Data Portal ingest) supplied in the prompt. That block returns zero AIA trial proceedings.
Verification searches run (2026-09-16): general web queries for "US4723149 PTAB inter partes review," "4,723,149" AIA trial PTAB petition, number-scoped IPR/CBM queries, Toshiba-owner-scoped validity queries, and the Unified Patents patent portal entry. None surfaced a petition, institution decision, FWD, or appeal. Caveat: I have no authenticated session on PTAB E2E / PatentCenter, so the statement below is a structured-data result corroborated by search, not a certified clearance. It is, however, about as safe as an absence can be — see Strategic summary for why.

Proceedings overview

There are no AIA trial proceedings of any type — 0 IPR, 0 PGR, 0 CBM — ever filed against US 4,723,149, and none can meaningfully be filed or are needed today, because the patent lapsed for non-payment of maintenance fees effective 1996-02-07 and its entire enforceable life closed out two decades before the AIA trial regime existed. The bottom-line defensive posture is not "the patent is hardened" and not "the claims were canceled" — it is "the patent is dead, so there is nothing to defend against." Any demand letter citing this number is facially deficient before you reach a single invalidity argument.

No proceeding to report

Because the structured source returned an empty set and no search corroborated a filing, there is no entry to populate with the requested per-proceeding fields (type, filed date, verbatim status, judge panel, grounds, institution decision, FWD claim-level verdict, settlement, appeal, defensive value). I will not manufacture a proceeding number — the prompt explicitly forbids it, and the record does not support one. Should a filing exist that the ODP ingest missed, the search terms above are the ones that would have surfaced it; they did not.

Strategic summary

Claim status: all five claims are UNTESTED at the PTAB — and untestable in any way that matters

No claim of 4,723,149 has ever been adjudicated by the Board. Claims 1–5 therefore stand as issued, unamended, with no PTAB-created prosecution history, no cancellation, no narrowing amendment, and no § 315(e) estoppel record. Ordinarily an untested five-claim patent with an unusually broad independent claim covering "detecting a secret document and erasing a user-selected area," and with 107 forward citations including the entire Digimarc/Canon anti-forgery lineage, would be prime IPR bait. That it has never been challenged is not evidence of strength. It is evidence of irrelevance: the patent lapsed 1996-02-07, and the AIA trial mechanisms (IPR from 2012-09-16; PGR for patents with pre-March-2013 priority, i.e. never available here; CBM sunset 2020-09-16) all post-date the patent's death. There was never a live enforcement campaign to provoke a petition.

The dispositive fact is the maintenance-fee lapse, not any invalidity ground

The Legal Events record in the patent file is unambiguous: maintenance-fee reminder mailed 1995-09-12; Lapse for failure to pay maintenance fees 1996-02-04; Lapsed due to failure to pay maintenance fee 1996-04-16, effective 1996-02-07; and a later status entry (2018-01-30) confirming "PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362." Under 35 U.S.C. § 41(b) and § 186, an unreinstated lapse terminates enforceability. Damages under § 286 are capped at six years pre-complaint — a window that closed in 2002 even if the lapse had never occurred. A demand letter on this number cannot support prospective relief (nothing is infringed after expiration) and cannot support backward relief (the lookback is long closed). Flag one reconciliation note for whoever is working the file: the Unified Patents portal (https://portal.unifiedpatents.com/patents/patent/US-4723149-A) shows priority 1985-05-07, filing 1986-05-04, grant 1988-02-01, and expiration 2006-05-04 — off-by-one-day artifacts already flagged in the earlier bibliographic section of this analysis. Even on Unified's most generous dates, the outcome is identical.

Pattern signals

• No serial petitioner, no serial patent owner, no aggregator. Unified Patents appears in the record only as a portal host, not as a petitioner — its page carries no litigation or trial entries for this number. There is no defensive-aggregator chain, no General Plastic / § 314(a) story, and no Director-level discretionary-denial practice to report.
• No patent-owner aggression at the Board. Toshiba (Kabushiki Kaisha Toshiba) never appealed anything concerning this patent to the Federal Circuit. The assignee of record has not asserted it. The 107 forward-citing documents cite it as prior art, which is the ordinary fate of an expired 1988 reference — not a sign of a live licensing program.
• Family check. The § 119 priority cases are JP 60-97313 (published as JPS61255362A) and JP 60-100918 (published as JPS61258278A). If someone is asserting "the 4,723,149 family," verify whether they have pivoted to a different, possibly still-live member. Do not accept a substituted number without checking the face of the document.

Estoppel landscape

Inapplicable. § 315(e)(2) estoppel binds a petitioner who obtained an FWD. There is no petitioner, no FWD, and no institution — so no prior-art ground is estopped against anyone. Conversely, there is no need for one. If a defendant were nonetheless inclined to build an IPR record for defensive symmetry (a waste of money here), nothing in the file would bar it.

Recommended next steps

1. If a demand letter cites US 4,723,149, do not brief validity. Respond with the maintenance-fee lapse: Legal Events, lapse effective 1996-02-07, and the 2018-01-30 discontinuation entry. State that the patent is unenforceable and that § 286's six-year damages window closed no later than 2002. There is no AIA proceeding to link to, because none exists — say that plainly rather than implying a validity win you never obtained.
2. Verify the number on the face of the pleading or letter. Tolling/assertion of a patent that lapsed in 1996 is a strong indicator of a mismatched, templated, or outright fraudulent demand, or of a scrivener's error pointing at a different right. Compare against the granted PDF and the § 119 priority documents above.
3. If any litigation is threatened, check the docket before assuming. I found no 2026 CAFC or district-court activity on this number, but I could not query PACER/CM-ECF or the CAFC docket. A live 2026 appeal or suit on a patent that lapsed in 1996 would be anomalous and should be confirmed in the case-management system if it is being represented to you as real.
4. If you must independently re-confirm the empty PTAB set, query the Board's E2E system by patent number (https://e2e.uspto.gov/, patent-number search) and cross-check CourtListener's docket/receipt search (https://www.courtlistener.com/?q=%224723149%22). Both are the correct venues to cite if you need a citable confirmation of "no proceedings." Because no FWD or CAFC opinion exists for this patent, there is no opinion link to provide — that absence is itself the answer.
5. Do not budget for an IPR defense. There is no asserted claim to invalidate, no estoppel to preserve, and no trial-stage milestone calendar (no institution deadline, oral hearing, or statutory one-year FWD date) to track, because no trial exists.

I'll research the assignment record for this patent before reconstructing the chain.

Let me search the assignment databases directly.

Let me verify the recorded reel/frame and check for any post-issuance transfers.

Research note before the findings

I was able to retrieve only a limited set of live sources before hitting my search-step limit: Google Patents (the authoritative full text supplied in this task), FreePatentsOnline, Espacenet, uspto.report, and the Unified Patents portal. I could not successfully load a raw USPTO Assignment Center / Assignments-on-the-Web record page for this patent; those queries returned empty result sets. Everything below that is marked as a recorded assignment therefore comes from the Google Patents legal-events feed embedded in the authoritative patent text, which renders the same underlying USPTO assignment records. I flag the one field I could not source (correspondent of record) rather than invent it.

Inventors

Unusual-pattern check — no departure pattern. The task flags "inventors departing the original assignee within 12 months of filing" as a fire-sale precursor. Here the movement runs the opposite way: Harada assigned into Toshiba approximately two weeks before the US filing (executed 1986-04-22; US application filed 1986-05-05), i.e., roughly 11.5 months after the JP priority date of 1985-05-08. This is a normal priority-establishing/employee assignment, not an inventor exit. There is a single inventor, so no co-inventor divergence can be assessed.

Original assignee

• Entity on the issued patent: Kabushiki Kaisha Toshiba (rendered "Toshiba Corp" on aggregator front pages), address as recorded: 72 Horikawa-cho, Saiwai-ku, Kawasaki, Japan. Assignment recorded 1986-05-05, effective 1986-04-22.
• Current assignee on record: Toshiba Corp — Google Patents shows no post-issuance assignee change, consistent with the single recorded conveyance below.
• Primary line of business: diversified Japanese electrical/electronics manufacturer (heavy electrical, consumer electronics, semiconductors, and office imaging). At the 1988 grant the patent's subject matter is an electronic copying machine — squarely within Toshiba's then-active copier line.
• Did they ship a product embodying the claims? Not determinable from the sources I could reach. Toshiba certainly sold copiers in this era, but I found no evidence tying a specific shipped model to the claim-1 combination (magnetic identification-section detection plus shared-carriage user-selected erasure). I will not assert embodiment without a product citation.
• Current status: Toshiba remained an operating listed company through the patent's life and long after. It was taken private by a consortium led by Japan Industrial Partners (JIP) and delisted from the Tokyo Stock Exchange in December 2023 (widely reported; I did not re-verify this in the limited searches available, so treat the delisting date as moderate-confidence background rather than a sourced finding here). No bankruptcy. Critically, none of that is relevant to this patent — it lapsed in 1996 and expired in 2006, two decades before the take-private.

Assignment timeline

There is one (1) recorded assignment in the chain. Google Patents renders the same reel/frame twice (once as the "owner name" field and once as free-format text) because it exposes both the structured and the OCR'd variants of the single record — that is one assignment, not two.

• 1986-04-22 (executed) / recorded 1986-05-05 — Reel 004551/0292
  • Conveyance: Assignment — recorded verbatim as "ASSIGNMENT OF ASSIGNORS INTEREST"
  • Assignor: Kentaro Harada (individual inventor)
  • Assignee: Kabushiki Kaisha Toshiba, 72 Horikawa-cho, Saiwai-ku, Kawasaki, Japan
  • Correspondent: Not retrievable from the sources I could reach. The Google Patents legal-events feed does not expose the assignment correspondent field; obtaining it requires the Assignments-on-the-Web abstract for reel 004551/0292 at https://assignmentcenter.uspto.gov/. I decline to name a correspondent rather than guess. (Note: the earlier-generated bibliography listed "Cushman, Darby & Cushman" as counsel of record. That is plausibly the prosecution attorney firm, which is a different field from the assignment recording correspondent, and I could not independently confirm it belongs to this reel/frame. Treat the two as distinct.)
  • Context: Initial employee/priority-establishing assignment to the original corporate assignee. Not a fire-sale, reorg, securitization, or transfer-to-asserter.

No post-issuance assignments are recorded. There is no assignment to any IP-holding entity, no security agreement, no merger, no change of name, and no release. The only later legal events on the record are administrative fee and status events, none of which are conveyances:

Reading of this finding: the absence of post-issuance assignments means Toshiba retained ownership of US 4,723,149 from the 1986 employee assignment until the patent lapsed in 1996 and expired on the anticipated 20-year date of 2006-05-05.

Timeline diagram

timeline
    title Ownership of US 4723149
    1985 : Inventor Harada files JP priority
    1986 : Harada assigns rights to Toshiba
         : Reel 004551 frame 0292 recorded
         : US application filed
    1988 : Patent issues to Toshiba
    1996 : Rights lapse for unpaid fees
    2006 : Anticipated 20 year expiry

NPE / troll-pattern signals

Verdict

Insufficient data — in the sense defined by the task as "no records, or only the original assignment." This patent has exactly one recorded conveyance (reel 004551/0292, executed 1986-04-22, recorded 1986-05-05, Harada → K.K. Toshiba), and no post-issuance transfers, licenses, or security interests.

Two things should be stated plainly alongside that label, because "insufficient data" here should not be read as "unknown risk":

1. As to NPE characterisation, the record is actually affirmative, not merely empty. All eight NPE-troll signals are not present, and the reason is structural: there is no downstream chain at all. The patent stayed with a large operating manufacturer from 1986 until it lapsed in 1996 (Toshiba's own record: "Lapsed due to failure to pay maintenance fee," effective 1996-02-07) and expired on 2006-05-05. There is no transfer-to-asserter to grade. The only gap is that a one-link chain offers nothing to test for the recurrence and cascading signals, and the correspondent-of-record field was not retrievable.
2. The single material open item is the correspondent field on reel 004551/0292, which I could not read from the sources available. It is a low-value item here — with one link from an individual inventor to Toshiba, a correspondent name cannot establish an NPE pattern either way.

Verification link: search patent number 4723149 at the USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (mirror: https://assignment.uspto.gov/patent/index.html) — and open reel 004551/0292 to read the correspondent field I could not retrieve.

Caveat I do not want buried: the negative findings above are bounded by what I could actually load. Free assignment feeds render US patents from the 1980s reliably, and the Google Patents legal-events feed (which I do have, in full, in the authoritative text) is itself derived from the USPTO assignment and fee records — so the "one assignment only" finding is well-grounded for this patent. What I could not do is query the raw USPTO abstract page directly or run a historical (1988–1996) litigation search; the prior litigation section already flagged that window as under-covered by free dockets.

I'll research the prior-art references of record for US 4,723,149, starting with the three U.S. patent citations and the IBM Technical Disclosure Bulletin.

Let me pin down the exact filing/grant dates for the two remaining U.S. citations and get the substance of the IBM TDB reference.

Prior-Art Analysis — US 4,723,149 (verified record)

Scope note. I matched the identifier exactly to US 4,723,149 / US4723149A (App. No. 06/860,007, filed 1986‑05‑05, priority JP 60‑97313 (1985‑05‑08) and JP 60‑100918 (1985‑05‑13), granted 1988‑02‑02). I did not substitute any similar number. I retrieved the record through public USPTO-derived mirrors (Google Patents, uspto.report, FreePatentsOnline, the granted-PDF image server) rather than a live authenticated PatentCenter session — so the bibliographic values below are quoted from those sources, not asserted as a certified USPTO extract. Builds on the earlier summary; not repeated here.

1. Legal framework applied

Because the application was filed 1986‑05‑05, pre‑AIA 35 U.S.C. § 102 governs (pre‑AIA § 102 applies to applications filed before 2013‑03‑16, and the '149 patent was never amended into AIA practice).

The operative § 102 dates for the '149 patent:

Critical point: the foreign priority dates (1985‑05‑08 / 1985‑05‑13) do not move the § 102(b) critical date. For pre‑AIA practice the one‑year grace runs from the U.S. filing date (1986‑05‑05), so anything publicly available on or before 1985‑05‑05 is a statutory bar. Conversely, a reference whose grant/publication falls after 1986‑05‑05 can only be reached under § 102(e), if at all.

2. The complete cited-reference set of record

The '149 front page lists three U.S. patent citations and two non‑patent citations (which are the same IBM article listed twice):

Note the shared-examiner fact: R. L. Moses was Primary Examiner on both US 3,977,785 and US 4,586,811, and is also the examiner credited on the '149 record. Both Xerox's and Ricoh's earlier cases were therefore in his own prior work.

3. Reference-by-reference analysis

A. US 3,977,785 — Harris, "Method and apparatus for inhibiting the operation of a copying machine"

Full citation: U.S. Patent No. 3,977,785; Inventor Ellis D. Harris; Assignee Xerox Corporation; App. Ser. No. 538,650; filed Jan. 6, 1975; granted Aug. 31, 1976; 22 claims, 9 figures. (Primary Examiner R. L. Moses.) Front-page dates per the granted PDF: https://patentimages.storage.googleapis.com/3d/16/53/8a31017d29f561/US3977785.pdf ; text at https://patents.google.com/patent/[US3977785A](/patent/US3977785A).
Portal discrepancy flagged (not corrected): Unified Patents lists priority 1975‑01‑05 / grant 1976‑08‑30 and Harvard-date artifacts — same off‑by‑one pattern already flagged for the '149 portal record. The granted document controls: Jan. 6, 1975 / Aug. 31, 1976.

Disclosure. A document carries a machine‑readable code (visible or invisible, e.g., near‑IR ink) printed at "any position on the document," and may be printed along the full width/length as a one‑dimensional line pattern. Light transmitted/reflected from the coded area is imaged through a reference transparency forming a correlation pattern on a photodetector array. If the document code correlates with the reference pattern, a "copy enable" signal is issued; if not, the copier is inhibited (the enable signal drives a relay in the machine's existing interlock chain, per its FIG. 7).

§ 102 status: granted Aug. 31, 1976 — nearly ten years before the '149 filing. It is § 102(b) statutory‑bar art (and § 102(a) art).

Element mapping to claim 1 (and dependents):

• "identification section" on the secret document → disclosed (its "code/signature," and stated to be locatable anywhere on the sheet ⇒ touches claim 3's "given position"; the width‑extending linear pattern ⇒ touches claim 2's "along one side").
• "secret document detecting means … sensor" → disclosed (photodetector array + reference transparency).
• "second controlling means … enabling or disabling the image forming operation" → disclosed (copy‑enable/inhibit logic).
• Missing: sensor "provided integrally with said original scanning means"; "two‑dimensional scanning" of the platen by the sensor; and every trimming/erasure element (light‑transmitting means, erasure‑portion specifying means, third/fourth controlling means, erasure storage, image erasing means).
• Not a magnetic system — so it does not reach claim 4.

Anticipation? No — it cannot anticipate claim 1, and therefore cannot anticipate dependent claims 2–5, each of which incorporates all of claim 1's limitations. It is citable against the detection half of claim 1 and as background.

B. US 4,586,811 — Kubo et al., "Confidential document reproduction prevention method"

Full citation: U.S. Patent No. 4,586,811; Assignee Ricoh Company, Ltd.; JP priority 1982‑04‑23; U.S. patent granted 1986‑05‑06; Primary Examiner Richard L. Moses; art unit classification 355/133. Text at https://patents.google.com/patent/US4586811 and https://uspto.report/patent/grant/4586811.
Exact U.S. filing date not confirmed in the sources I retrieved — I could not pull the § 102(e) filing date for this reference before hitting my research limits. The JP priority (1982‑04‑23) and the 1986 grant make a mid‑1983 U.S. filing overwhelmingly likely, but I am flagging this as unverified rather than asserting it.

Disclosure. The claims are drawn to a method: attach a rare‑earth‑doped fluorescent material (a long enumerated list — YF₃, LaF₃, GdF₃, BaYF₅, Y₂O₂S, Y₃OCl₇, etc., doped with Ce/Yb plus Er/Ho/Tm) to the confidential original; in the reproduction machine, expose the original to infrared so the material emits visible light; detect that emission with a photosensor; and stop or cancel at least one step indispensable for reproduction in response to the detection signal. The specification shows several "confidential mark detectors," including scanning‑mounted detectors, detector arrays, and facsimile‑machine implementations.

§ 102 status — this is the key date trap: the patent issued May 6, 1986, which is one day after the '149 U.S. filing date of 1986‑05‑05. It is therefore not § 102(a) art and not § 102(b) art as a publication. Its only viable § 102 footing against the '149 claims is § 102(e) ("described in a patent granted on an application for patent by another filed in the United States before the invention thereof by the applicant"), which depends on the unverified U.S. filing date being earlier than the '149 invention date. If that U.S. filing date is not before the '149 invention date, US 4,586,811 is not § 102 art against the '149 patent at all.

Element mapping to claim 1:

• "identification section" / detecting means / "disabling the image forming operation" → disclosed in substance (fluorescent confidential mark; IR excitation; visible‑light detection; reproduction step cancelled).
• Missing: "provided integrally with said original scanning means"; two‑dimensional sensor scanning of the platen; and the entire trimming/erasure half of claim 1. Its mark does not appear to be a magnetic medium (claim 4).

Anticipation? No as to claims 1–5. Even if § 102(e) status is established, it discloses fewer of claim 1's limitations than US 3,977,785 does, and none of the erasure limitations.

C. US 4,486,090 — "Copy machine with unauthorized copy prevention"

Full citation: U.S. Patent No. 4,486,090; Assignee Minnesota Mining and Manufacturing Company (3M); U.S. filing 1983‑03‑21, Ser. No. 477,061; granted Dec. 4, 1984. Corresponding EP 0 122 729 B1 (copy machine with document dependent control). Text at https://patents.google.com/patent/[US4486090A](/patent/US4486090A)/en ; granted PDF at https://patentimages.storage.googleapis.com/61/d5/79/041e48bff987d5/US4486090.pdf ; EP family data showing "US 47706183 A 19830321" at http://data.epo.org/gpi/EP0122729B1-COPY-MACHINE-WITH-DOCUMENT-DEPENDENT-CONTROL.html.

Disclosure. An eddy‑current proximity detector is mounted on the scanning carriage so that relative movement arises between detector and original when the copier runs; when metal is carried by the document the detector output trips a latching circuit, and control logic inhibits the charge corona and transfer corona so no copy can be made. Additional aspects: a document‑cover position detector required to be closed before a copy cycle; a timing reference (position indicator plus pulse generator) so the control logic knows when the detector is over the document region; and a self‑check in which a metal target in the scan path verifies detector operability.

§ 102 status: granted Dec. 4, 1984 — plainly before the 1985‑05‑05 critical date. Solid § 102(b) art (also § 102(a)).

Element mapping to claim 1 (and dependents):

• "original scanning means, movable along said original table" → disclosed (carriage 5 on shafts 10/11 with reciprocating drive 12).
• "secret document detecting means provided integrally with said original scanning means" → disclosed almost verbatim: the detector is placed on the scanning apparatus and detection occurs through the scan motion ("placing the detector on the scanning apparatus"; "relative movement … is provided after the document … is positioned on the platen"). This is the closest of the four to that express limitation.
• "first controlling means for driving said document scanning means prior to an image forming operation" → partially disclosed (detection occurs during carriage travel before image capture; position indicator/pulse generator give the pre‑capture timing) — though it is not a separate pre‑scan pass.
• "second controlling means … disabling the image forming operation" → disclosed (inhibit charge/transfer corona; latch until reset).
• Touches claim 4's concept (metal/magnetic‑field detection) but not literally: it senses eddy currents in metal, not a "magnetic medium" read by a "magnetic sensor." Claim 4 requires the pairing of magnetic medium + magnetic sensor; a literal reading is not met.
• Missing: two‑dimensional (x‑and‑y) sensor scanning of the platen; and the whole trimming/erasure half of claim 1.

Anticipation? No as to claims 1–5. It is the strongest of the three U.S. citations on the "integral with the scanning means" + "disable copying" limitations, and would be a lead § 103 reference, but it is not a single‑reference anticipation of any claim.

D. IBM Technical Disclosure Bulletin — Guido, "Preventing Copying of Classified Information"

Full citation: A.A. Guido, "Preventing Copying of Classified Information," IBM Technical Disclosure Bulletin, Vol. 19, No. 4, September 1976, pages 1469–1470. Listed twice in the '149 record (the two duplicate non‑patent‑citation entries).
Corroborating appearances: cited in the International Search Report for WO 1985/001128 A1 (https://patentimages.storage.googleapis.com/ea/05/a3/df3780147f28d0/WO1985001128A1.pdf) and in US 4,908,873 ("Document reproduction security system," https://patents.google.com/patent/[US4908873A](/patent/US4908873A)/en).

Disclosure — honesty flag: I did not retrieve the two‑page article text itself; the record above is limited to confirming its identity, volume/issue/page, author, and date from the citing documents. Based on those citations and on the surrounding IBM TDB family for copier document detection (e.g., the magnetic‑field generator/detector‑coil and conductive/metal‑backed‑original approaches described in US 4,486,090's background), the Guido disclosure is a copier‑disable arrangement that senses a special physical characteristic of the original (magnetic/metallic) and inhibits reproduction of "classified" documents generally — i.e., generic genus‑level teaching, not a specific two‑axis sensor architecture. I would want the actual pages before attributing any specific circuit or sensor geometry to it.

§ 102 status: published September 1976 — a printed publication under § 102(b), far more than one year before the '149 filing date.

Element mapping: genus‑level only — it supports "secret document" + "detection" + "disable copying" as old and known, and is cumulative of the "identification section is a magnetic medium / sensor is a magnetic sensor" idea behind claim 4. It discloses none of claim 1's scanning‑integration, two‑dimensional‑scan, or erasure limitations.

Anticipation? No as to claims 1–5. Citable for the claim 4 concept and as § 103 background only.

4. Anticipation matrix (§ 102)

Bottom line on § 102: None of the four cited references anticipates claim 1, and because claims 2–5 all depend from claim 1 and incorporate its erasure/area‑selection limitations, none anticipates claims 2–5 either. Every cited reference addresses only the secret‑document‑detection half of claim 1. Not one discloses the second half that claim 1 requires as an integrated whole:

• light‑transmitting means (spot‑light source 171) integral with the scanning means;
• erasure‑portion specifying means (keys 30a–30g defining a rectangle by two spotted points, or its inverse);
• third controlling means driving the scanning means to the commanded spot position;
• erasure‑portion storage (RAM 140 bitmap);
• image erasing means (LED erasure array 150 between charger 11 and exposure region Ph, or between Ph and developer 12); and
• fourth controlling means reading the stored data out during imaging (shift/store registers 161/162 + switch array 164, latched within θ₁/ω).

The single‑reference rule of § 102 therefore defeats anticipation across the board. The references are properly § 103 material (and, for US 3,977,785 / US 4,486,090 / the IBM TDB item, also § 102(b) material against individual limitations), which is consistent with why the '149 patent issued over them: the asserted inventive contribution was not "detect a marked original and stop the copier" — that was old by 1976 — but sharing one two‑axis positioning system on the scanning carriage between the security sensor and the trimming spot‑lighter, as express claim language now reflects ("provided integrally with said original scanning means" in both the detecting‑means and light‑transmitting‑means elements).

5. Caveats and open items

1. Examiner's actual rejections are not in the extracted record. The bit‑map above is derived from the cited references and claim text, not from the prosecution history. Whether the examiner applied these references under § 102, § 103, or merely as Form SB/08 background cannot be determined from the Google Patents rendering; the file wrapper would be needed. Treat the "§ 102 basis available" column as statutory availability, not as a record of an actual rejection.
2. US 4,586,811's U.S. filing date is unverified. Its § 102(e) reach against the '149 claims is contingent on that date preceding the '149 invention date. I flagged rather than assumed this.
3. The IBM TDB article text was not retrieved. Its characterization above is reconstructed from citing documents; I expressly decline to attribute specific content to pages 1469–1470 that I have not read.
4. Do not read "no anticipation" as "weak patent." Anticipation and validity are different questions; the four references are relevant, material art and the three U.S. citations plus the IBM item would almost certainly figure in any § 103 challenge to the detection sub‑combination.
5. Similar‑number artifacts excluded. Search hits pointing to other Toshiba patents (e.g., the 4,727,513 family) and to unrelated 4,7xx,xxx numbers were identified and excluded; only exact‑match US 4,723,149 material is reported.
6. The earlier-noted off‑by‑one portal discrepancies (Unified Patents) recur on US 3,977,785's record; in both instances the granted document controls and I did not auto‑correct either set.

I'll verify the prior-art references of record before building the § 103 analysis.

Strong hit on the Ricoh reference. Now let me pull the 3M patent, the IBM TDB, and the area-selection/erasure art.

Now the area-selection/erasure art and the IBM TDB content.

Obviousness Analysis — US 4,723,149 under 35 U.S.C. § 103

Note on dates: This task states the current date as April 26, 2026; the session header for this matter reads 2026-09-16. Immaterial to the analysis below (all math runs off the 1986-05-05 filing date), but flagged rather than silently reconciled.

1. Governing framework and the threshold I have to clear first: which references are actually available

The application was filed 1986-05-05, so pre-AIA §§ 102/103 apply. That fixes three effective cut-offs, and they matter enormously here because most of the material on the Google Patents page is not § 103 prior art:

This eliminates most of the "Similar Documents" list. The Toshiba area-selection family that supplies the trimming/erasure half of claim 1 — US 4,666,288 (JP priority 1985-01-18, US filed 1986-01-17, granted 1987-05-19), US 4,653,899, US 4,655,580, EP 0 180 984, US 4,884,107 (1989) — is chronologically marginal to unavailable: an 18-month JP kokai of the 1985-01-18 filing would publish ~1986-07-18, after both cut-offs, and EP 0 180 984's publication date sits right at the 1986-05-05 filing date (I did not verify it — it must be checked in the EPO Register before anyone relies on it). Worse, to the extent any of that family qualifies only under § 102(e), pre-AIA § 103(c) excludes it from an obviousness combination because it was commonly owned by Toshiba at the time the '149 invention was made. In other words: the prior art that discloses the erasure half of claim 1 is largely the patentee's own later work and is statutorily walled off. That is the single biggest structural fact in this analysis.

Genuinely available art:

Also surfaced but outside the record and unverified as to filing date: US 4,603,970 (granted 1986-08-05), which states "The scanning mechanism of the sensor 4 may be shared with the optical scanning mechanism of, e.g., an exposure lamp." If its US filing predates May 1985 it is § 102(e) art and is a direct motivation-to-combine quote. Verify before use.

2. Element mapping — claim 1

3. The combinations that would render the claims obvious

Ground A — Claims 2, 4, 5 (detection sub-combination): strong.
Any one of '785, '811, or '090 alone, or '811+'090, discloses the detection-and-disable subject matter of claims 2–5.

• Claim 4 (magnetic medium + magnetic sensor): '090 uses magnetic-field-adjacent eddy-current sensing of metal on the document; '785's background expressly enumerates "magnetic" among recognized code types and cites IBM TDB 15(12) (May 1973) magnetic-flux detection. Substituting a magnetic tape strip for a metal flag/coating is a predictable substitution of known detection modalities for their known function (KSR, In re ICON). Obvious.
• Claim 2 (mark along one side): '090's metal flag/backing and '811's Fig. 3a–3c mark arrangements across the document teach edge placement as a design choice.
• Claim 5 (start from the position farthest from the reference scale): '090's pulse-generator/position-indicator scheme establishes known carriage position and a defined start reference; '811 and '785 likewise need a known scan origin. Selecting the far end as the origin so the detection sweep terminates at the reference scale is a routine scan-sequencing choice with the predictable benefit of a single sweep.

Ground B — Claim 1: obvious only if an available reference supplies the trimming half. The KSR-rational combination would be:

Primary: '811 or '090 (carriage-mounted detector + pre-scan detection + copy disable)
+ Secondary: area-erasure/trimming system (see-through spot light on a transverse sub-carriage, pulse-counted coordinates, RAM bit-map, LED discharge array with shift/store register and latch timing)

Motivation (KSR factors):

1. Common hardware / duplication avoidance. Both functions require the same thing: positioning a small head in x and y over the platen. '811 already puts a sensor on the optics carriage; '090 expressly teaches putting the detector on the scanning apparatus. Mounting the mark sensor on the existing spot-light sub-carriage reuses its stepping motor, timing belt, guide shaft, and home sensor — the '149's own specification concedes this is the whole point ("a convey means… can be used in common… a simple arrangement").
2. Both reference families are in the same field and address the same problem (unauthorized reproduction of selected originals) and both concern the same machine architecture.
3. Predictable result, no change in principle of operation. Each element performs exactly its known function; the combination yields no unexpected interaction.
4. Design incentive at the relevant time. A copier designer told to add copy protection to a machine that already offered area erasure would naturally use the already-present x-y positioning system rather than add a second one.

Ground C — the "prior to image forming" element, if attacked separately: '785's two-pass architecture (region-b detection scan, then region-a copy scan only if a copy-enable signal appears) is an explicit pre-copy detection teaching; '811's "before starting copying" is even more explicit.

4. Where the obviousness case breaks down — and it does break down

1. Two-dimensional detection is the hinge element, and nothing in the available art teaches it. '090's coil is a single carriage-mounted point sensor; '811's photosensor moves with the lamp mirror. Neither performs an x-and-y raster over the platen. The '149's stated reason for the raster — detecting the mark anywhere on the page, including mid-document (claim 3), regardless of how the user placed the original — is a distinct functional goal, and the record contains no motivation to sweep in the transverse axis for detection (only for trimming, which is Toshiba's own art).
2. The entire erasure/trimming half of claim 1 is unsupported by available art. See § 1. Toshiba's sibling applications supply it almost verbatim (the EP 0 180 984 specification, for instance, reproduces the '149's erasure-array and θ₁/ω latch paragraphs nearly word-for-word), but they are the patentee's own work, published at or after the filing date, and excluded from a § 103 combination by § 103(c) to the extent they are § 102(e) art. An examiner combining '811 with Toshiba's own contemporaneous disclosure would have had to ignore that the secondary subject matter was the same applicant's — a ground the applicant would defeat on § 103(c).
3. Express teaching away. '785 states that code-recognition copy-prevention "would be that the system be complex and expensive to make" and frames its own contribution as avoiding sensor-based code recognition in favor of static optical correlation. A POSITA reading '785 is being pushed toward a fixed reference transparency, not toward a moving magnetic sensor.
4. Prosecution history. The three cited references plus the IBM TDB were on the record, and the claims issued. That is evidence, not holding, that the examiner did not view the combination as prima facie obvious — and the references cited are all detection-side art, consistent with my own reading.

Secondary considerations: the 107 forward citations (Canon's 1988 anti-forgery family, the Digimarc watermark families) mark the '149 as a starting point for later document-security work, but forward citations are not themselves secondary considerations; no evidence of nexus, commercial success, licensing, or long-felt-need-but-unsolved appeared in the record I reviewed. The maintenance-fee lapse in 1996 is weak evidence against commercial significance but proves little on its own.

5. Verdict

Bottom line: the claims most exposed to § 103 are 2, 4, and 5. Claim 1 would survive the record's own prior art, and the reason is not the abstract idea of "detect-and-inhibit" (that is squarely old) but the specific architecture — a two-dimensionally swept platen raster executed by a security sensor that shares a transverse sub-carriage with a see-through trimming light — plus the chronology/ownership accident that hides the enabling trimming art behind § 103(c).

Flagged uncertainties

• IBM TDB (Guido) content: not verified. I found the citation in three independent places but no body text. If it in fact discloses a magnetically-coded original detected before copying and a machine disable, it materially strengthens Ground A/B for claims 1–4 — verify it (the volume is available in most corporate libraries and via the IP.com TDB archive).
• EP 0 180 984 publication date: not verified. It is within days of the '149 filing date. If it published on or before 1986-05-05, the analysis in § 4.2 changes substantially (subject to § 103(c)).
• US 4,603,970 filing date: not verified; if pre-May-1985 it supplies an express motivation to share the sensor's scanning mechanism with the exposure-lamp scanner.
• No independent (non-Toshiba) pre-May-1985 area-erasure art was located through the searches run; I did not exhaustively search the G03G15/36 subclass and a paid database search could well find Canon or Xerox trimming art that closes the § 4.2 gap.
• Minor: Unified Patents' portal gives US 3,977,785 as priority 1975-01-05 / granted 1976-08-30, while the granted front page reads filed Jan. 6, 1975 / patented Aug. 31, 1976 (same off-by-one pattern flagged for the '149 itself in the earlier section). The granted document controls.

Sources: US 3,977,785 PDF · US 4,586,811 full text · US 4,486,090 · US 4,666,288 · EP 0 180 984 · Guido TDB citation · US 4,603,970