US 4806774

I'll run several searches to gather authoritative information on US 4806774, including any litigation activity.

Let me search for the actual claim text and any litigation/docket activity.

US 4,806,774 — Verification Summary

Important caveat up front: I was able to confirm bibliographic data, the abstract, and the specification, but the claim text itself was not retrievable in the material available to me (the patent copy provided cuts off mid-way through the detailed description, in the Time Delay Integration section, before the claims). The claim overview below is therefore derived from the specification and from the EP family member (EP 0 294 643 B1), which shares the same disclosure, and is flagged as such. I also found no CAFC (or any court) docket activity for this patent — see the litigation section.

---

Bibliographic Data (confirmed)

Source: https://patents.google.com/patent/US4806774/en and https://patents.google.com/patent/USRE33956F1/en

---

Abstract (as published)

An inspection system (10, 100) employs a Fourier transform lens (34, 120) and an inverse Fourier transform lens (54, 142) positioned along an optic axis (48, 144) to produce from an illuminated area of a patterned specimen wafer (12) a spatial frequency spectrum whose frequency components can be selectively filtered to produce an image pattern of defects in the illuminated area of the wafer. Depending on the optical component configuration of the inspection system, the filtering can be accomplished by a spatial filter of either the transmissive (50) or reflective (102) type. The lenses collect light diffracted by a wafer die (14) aligned with the optic axis and light diffracted by other wafer dies proximately located to such die. The inspection system is useful for inspecting only dies having many redundant circuit patterns. The filtered image strikes the surface of a two-dimensional photodetector array (58) which detects the presence of light corresponding to defects in only the illuminated on-axis wafer die. Inspection of all possible defects in the portions of the wafer surface having many redundant circuit patterns is accomplished by mounting the wafer onto a two-dimensional translation stage and moving the stage (40) so that the illuminated area continuously scans across the wafer surface from die to die until the desired portions of the wafer surface have been illuminated. The use of a time delay integration technique permits continuous stage movement and inspection of the wafer surface in a raster scan fashion.

---

Plain-Language Overview of the Independent Claims

Confidence level: moderate. The exact claim wording was not in my source material. The following is reconstructed from the disclosure, the official abstract, and the European counterpart (EP 0 294 643 B1), which states: "in one aspect, the invention provides a detection method having the features of claim 1. In another aspect, the invention provides a detection system having the features of claim 19." This indicates the US patent is similarly structured with at least one method independent claim and one system independent claim. The EP search report cited claims 1–4, 6, 7, 10–20, 24 and 28 against the prior art, suggesting the US counterpart has at least 28 claims.

Independent Claim 1 — Method (paraphrase):
A method of inspecting a specimen (e.g., a semiconductor wafer) whose surface carries an array of dies, each die containing many redundant circuit patterns. The method steps are:

1. Illuminate a region of the patterned specimen surface with collimated coherent (laser) light — the illuminated region is large relative to the imaged die, so the Fourier transform is built from many repetitions of the redundant pattern.
2. Pass light diffracted by the illuminated region through a Fourier transform lens to form a spatial frequency spectrum (the intra-die diffraction pattern) in the back focal plane.
3. Place a pre-fabricated spatial filter in that plane that blocks/absorbs (or scatters) the spatial frequency components of the error-free Fourier transform, while passing light from defects.
4. Perform an inverse Fourier transform on the filtered light with a second lens to form a defect image.
5. Detect that image with a two-dimensional photodetector array positioned centrally on the optic axis, with a photosensitive area too small to cover the whole image field, so that only the on-axis die's defects are captured.
6. Scan the specimen with a two-dimensional translation stage so that successive portions of the die/wafer are moved through the illuminated on-axis region, preferably continuously, in stripe-by-stripe raster fashion using time delay integration (TDI).

The asserted novelty/point of departure: prior systems (Watkins US 4,000,949; Mathisen US 3,614,232) required both inter-die and intra-die interference information; the '774 invention deliberately uses intra-die information only, restricted to dies having many redundant circuit patterns, which permits a small-object-field lens and an essentially aberration-free transform.

Independent Claim 19 (or thereabouts) — System/Apparatus (paraphrase):
An inspection system comprising the same combination in apparatus form: a laser light source producing a collimated beam; a Fourier transform lens receiving light diffracted from an illuminated area of the patterned specimen; a spatial filter located in the Fourier transform (back focal) plane that blocks the spatial frequencies of the error-free Fourier transform and passes defect light; an inverse Fourier transform lens; a two-dimensional photodetector array (e.g., a charge-coupled device) centered on the optic axis in the image plane with a light-sensitive surface smaller than the image field; and a two-dimensional translation stage (X-Y table, e.g., Kensington Laboratories Model 8500) for scanning the specimen through the on-axis illumination region, operated continuously to implement time-delay-integration.

Disclosed refinements reflected in dependent claims (likely subject matter): the transmissive spatial filter 50 vs. the reflective liquid-crystal (laser smectic light valve) spatial filter 102; the folded polarization-based optical configuration (polarizing beam splitter 116, quarter-wave plates 118/130); in-situ fabrication of the filter by exposure to light diffracted by all dies; the specific 10X magnification / 3 mm object field / 20 mm illumination spot geometry; and the specific 403 × 512 element CCD raster-scan TDI detection scheme.

---

Litigation / CAFC 2026 Dockets

No results. I searched specifically for "4806774," "US4806774," and the reissue "RE33956" in connection with CAFC / Federal Circuit / 2026 docket activity and found nothing. Given that the patent expired June 8, 2007, and both the original and reissue are marked "Ceased," active 2026 appellate litigation is highly unlikely. I cannot state with certainty that no case ever existed, but I found no authoritative record of one.

If you need a definitive litigation check, the authoritative sources would be the USPTO Patent Center / Assignment records for 4806774 and PACER / the Federal Circuit docket (CM/ECF) for party-name searches such as "Insystems" or "Optical Specialties."

---

Prior Art of Record

• US 4,000,949 (Watkins) and US 3,614,232 (Mathisen) — the two systems expressly distinguished in the Background; both are described as requiring inter-die + intra-die interference information and simultaneously inspecting all dies.
• European search report additionally cited EP-A-0 017 371 (Tokyo Shibaura Denki K.K.), a Xerox Disclosure Journal article by W. M. Sterling et al. ("Reticle inspection using diffraction pattern masking"), and US-A-4,370,024 (H. L. Task et al.). The EP B1 specification also discusses an Iwamoto system that blocks higher diffraction orders and scans large samples.
• The patent is cited by later Insystems work, e.g., US 5,428,442 ("Inspection system with in-lens, off-axis illuminator") and US 5,172,000 ("Spatial filter for optically based defect inspection system").
• Kahn, Frederic J., "Large Area, Engineering Drawing Quality Displays Using Laser Addressed Smectic Liquid Crystal Light Valves," Automation Technology Institute Conference, Montreal, Canada, February 1987 — cited as describing the light-valve type used for spatial filter 102.

---

Explicit Uncertainties

1. Exact claim language is not confirmed. The claim set was not present in my material. The paraphrases above are reconstructions and should not be quoted as the literal claims. Verify against the granted patent PDF (USPTO PatentCenter / Google Patents PDF) or the reissue RE33956, which may have amended claims.
2. Independent-claim numbering ("claim 1 method," "claim 19 system") is inferred from the EP 0 294 643 B1 statement of aspects; I have not confirmed the US claim numbering.
3. Assignee spelling — "Optical Specialities, Inc." appears as given in the record; I am not normalizing it to "Specialties."
4. No litigation found is not proof of no litigation.

I'll search for litigation involving US patent 4806774.

Let me search additional litigation-specific sources.

Let me run more targeted searches on litigation databases.

Litigation Search Report — U.S. Patent No. 4,806,774

Patent identification (as literally identified): U.S. Patent No. 4,806,774 — "Inspection system for array of microcircuit dies having redundant circuit patterns." Application No. 07/060,090; filed June 8, 1987; granted/issued February 21, 1989; inventors Lawrence H. Lin, Daniel L. Cavan, Robert B. Howe; original assignee Insystems, Inc.; listed current assignee Optical Specialities, Inc. Term anticipated to expire June 8, 2007 (status: Ceased).

Finding: No litigation identified

Based on my searches, I found no record of any litigation (district court, CAFC, or ITC) involving U.S. Patent No. 4,806,774. I am not able to identify any plaintiff, defendant, jurisdiction, case number, filing date, or outcome, because no such case surfaced in the sources I was able to search.

What I was able to verify

Instead of litigation, the only "case-like" paper trail for the '774 patent is its USPTO assignment/security-interest history, which does not evidence enforcement activity:

• 1988-04-12 — Assignment to Insystems, Inc. (inventors Cavan, Howe, Lin).
• 1992-11-19 — Security interest granted to Silicon Valley Bank (assignor Optical Specialties, Inc.) — a financing lien, not litigation.
• 1993-11-23 — Assignment of interest from Insystems to Optical Specialities, Inc.
• 2007-06-08 — Anticipated expiration; legal status currently listed as Ceased.

This assignment chain (startup → bank collateral → successor entity → expiration) is consistent with a patent that was never asserted in court.

Important caveats on methodology

I want to be explicit about the limits of this conclusion rather than overstate it:

1. Search coverage was incomplete. My searches of general web sources, Google Patents, and Unified Patents-related material did not return results from a litigation-specific database (Unified Patents litigation portal case list, RPX/Insight, PACER, or the CAFC docket) for this specific patent number. My access to those paywalled/portal-specific databases was limited, and my final targeted queries were cut off before completion.
2. A negative result is not proof of a negative. Given the patent's age (issued 1989, expired 2007), any litigation would predate modern docket-aggregation coverage, and pre-2000 suits may not be fully indexed in free databases. The absence of hits is therefore weak evidence of no litigation, not conclusive.
3. No similar-number results were substituted. I excluded hits for other patents (e.g., 4,806,774-adjacent numbers, the '710 patent in Bott v. Four Star, the '677/'508 patent cases, etc.), none of which involve U.S. 4,806,774.

Recommended next steps to confirm

To reach a defensible conclusion, the following sources should be checked directly (which I could not fully query here):

• Unified Patents Litigation Portal — https://portal.unifiedpatents.com/litigation/caselist (search by patent number).
• RPX Insight — https://insight.rpxcorp.com (patent-level litigation tab).
• PACER / CourtListener RECAP — docket search for "4,806,774" and "4806774," including the 1990s districts where Insystems/Optical Specialties were active (e.g., N.D. Cal.).
• CAFC docket and opinions — search for appeals naming Insystems or Optical Specialties.
• USPTO Patent Assignment / litigation records and the Patent Trial and Appeal Board (no IPR/PGR is possible in this era, but reexaminations could be checked).

Bottom line: As of this search, there is no known litigation involving U.S. Patent No. 4,806,774, and the available ownership record suggests the patent was never asserted. However, this should be confirmed against a dedicated litigation database before being treated as a definitive negative.

I'll verify the structured data against public sources before writing up.

Proceedings overview

Total AIA trial proceedings on US 4,806,774: 0 — the USPTO ODP "PTAB proceedings on file" block is canonical and reports no AIA trials, and my independent web searches surfaced no IPR, PGR, or CBM petition, no Final Written Decision, and no Federal Circuit appeal naming this patent (breakdown: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denied). The bottom-line defensive posture is therefore not "hardened by surviving IPRs" and not "claims canceled by the Board" — it is something stronger for a defendant: the patent expired on 2007-06-08, roughly five years before the AIA created IPR/PGR/CBM at all, so no AIA challenge was ever possible in practice, and any 2026 demand letter citing US 4,806,774 is asserting an expired patent whose claims were surrendered and replaced by a reissue in 1992.

No proceedings to report

I will not manufacture docket numbers to populate the per-proceeding template. Instead, here is why the count is zero, dimension by dimension, since the reason matters to the defense:

• PGR — structurally impossible. Post-grant review is available only for patents with a filing date on or after 2013-03-16. US 4,806,774 was filed 1987-06-08 (priority 1987-06-08) and granted 1989-02-21 (Google Patents).
• CBM — ineligible and sunset. The transitional covered-business-method program ran 2012-09-16 to 2020-09-16 and applied only to patents claiming a financial product or service. This patent claims an optical wafer/photomask defect-inspection system (Fourier transform lens + spatial filter + inverse transform lens + photodetector array). It is not a covered business method.
• IPR — never filed. Expiry alone is not an absolute jurisdictional bar to IPR (the Board has instituted IPRs on expired patents, where an unexpired claim of the same family or pre-expiration damages exposure sustains the petitioner's interest), but no petition exists on the record. Practically, the patent expired 2007-06-08 — i.e., before the AIA took effect — and Google Patents lists the legal status as "Ceased" with an "Anticipated expiration" of 2007-06-08, consistent with a lapse for non-payment of maintenance fees.
• Reissue overlay — the claims you would attack no longer exist in their original form. The file history shows a reissue application (US07/613,208, priority 1990-11-14) issuing as US RE33956 E, granted 1992-06-09. Under 35 U.S.C. § 252 the original claims are surrendered upon reissue. Note that US 4,806,774 is itself cited as prior art against later third-party patents — e.g., in the ISR for WO 2013/114022 and in EP 1 231 460 A3, and it is the acknowledged background in US 5,172,000 ("U.S. Pat. No. 4,806,774 of Lin et al. describes an optical inspection system…"), so its disclosure is firmly in the public domain prior-art corpus.
• Ownership: originally Insystems Inc.; assigned to Optical Specialties, Inc. (recorded 1993-11-23), with a Silicon Valley Bank security interest recorded 1992-11-19. Current assignee of record listed as Optical Specialities Inc. No defensive aggregator (e.g., Unified Patents) appears anywhere in the assignment chain — searches surfaced none.

---

Strategic summary

Claim status. No claim of US 4,806,774 has ever been canceled, confirmed, or construed by the PTAB, because no AIA proceeding was ever filed. The operative narrowing event for this patent is not an IPR but the 1992 reissue (RE33956, granted 1992-06-09), which surrendered the originally issued claims and replaced them with the reissue claims. The reissue's term is the unexpired remainder of the original term, and both the original patent and the reissue are now long expired. I could not verify the specific claim set of the '774 patent from the text available to me (the supplied record breaks off mid-Detailed Description, before the claims), so I am not going to assign claim numbers a fate they never had. What I can say with high confidence: there is no surviving, enforceable claim of this patent in 2026, and therefore no "surviving claims" list to hand you.

Estoppel landscape. Because there is no IPR/PGR/CBM anywhere on this patent, § 315(e)(2) estoppel is a non-issue in both directions — no petitioner is barred, no patent owner is protected by a prior Board win, and no IPR record exists to be used offensively (no FWD findings, no APJ panel, no claim-construction record to cite). Every prior-art ground is formally "available," but that is a hollow victory: the real defenses are far cheaper and dispositive — (i) the patent expired 2007-06-08, so there can be no ongoing or future infringement; (ii) the § 286 six-year damages lookback means recoverable damages, if any, would reach only conduct in the six years before suit — i.e., none, since 2007; (iii) with no post-expiration infringement, injunctive relief is unavailable as a matter of law; and (iv) US RE33956 and the expired '774 patent's own claims must be traced against the accused product, which is a § 252 / reissue-claim question the patent owner would have to brief from a 1992 record.

Pattern signals. None of the usual troll patterns are present: no repeat petitioner filers (zero petitioners), no patent-owner appeals to the Federal Circuit on PTAB outcomes (zero FWDs to appeal), no reexamination or reissue-contested history visible beyond the 1992 reissue itself, and no aggregator involvement in the chain of title. The foreign family — EP 0 294 643 B1 / DE 3888395 T2 (both filed 1988-05-24), JP 2639501 B2 (filed 1988-06-08), KR 0132435 B1 (filed 1988-06-08) — was filed contemporaneously with the US case and would also have reached end of term around 2008; I recommend confirming each in its national register before relying on that, as I did not verify them individually. A patent that expired nineteen years ago and attracted no AIA petition is not a neglected-but-live asset; it is an abandoned one.

Recommended next steps

1. Do not file an IPR. There is nothing to institute against. There is no FWD to link to, no disposition to quote, and filing a petition on a patent that lapsed 2007-06-08 would be a waste of the filing fee. If a demand letter arrives, respond on expiry and § 286 grounds rather than commencing an AIA trial.
2. Verify and document expiry, then paper the record. Pull the maintenance-fee history and the "Ceased" status from USPTO Patent Center (patent number 04806774) and the Patent Assignment records, and confirm whether RE33956 was maintained through term. Confirm the expiry date you intend to rely on — the pre-URAA rule gives the longer of 17 years from issuance (2006-02-21) or 20 years from filing (2007-06-08), which is why 2007-06-08 appears as the anticipated expiration.
3. Check the family, not just the US patent. If the demand letter gestures at "the patent family," confirm the lapsed/expired status of EP 0 294 643 B1 (incl. DE 3888395 T2), JP 2639501 B2, and KR 0132435 B1 in their national registers before conceding anything about foreign exposure.
4. If assertion continues despite expiry, the escalation path is a declaratory-relief action for non-infringement/invalidity or an early motion attacking the complaint's failure to plead post-expiration infringement — plus a Rule 11 / § 285 fee position if the letter affirmatively cites claims of a patent that expired in 2007. That is where a fee-shifting threat has real teeth; an IPR threat here does not.
5. Update the docket on any change. If a petition is somehow filed now, it would appear on PTAB E2E; the statutory clock would then run one year from institution to FWD under § 316(a)(11). As of the 2026-09-30 ODP ingest, that has not happened, and I found nothing in web sources contradicting the ODP's "no AIA trial proceedings" result.

I'll research the assignment record for US 4806774 across the sources specified.

Inventors

Departure pattern — not present. The "all inventors gone within 12 months" precursor does not appear:

• Lin and Howe are still at Insystems on the next family member, US 5,172,000 ("Spatial filter for optically based defect inspection system"), filed 1990-11-02, granted 1992-12-15, assignee Insystems, Inc. (per Justia's "Patents Assigned to Insystems, Inc." listing).
• Cavan, Lin and Howe all co-published from Insystems in November 1988, ~17 months after the June 1987 filing (JVST B 6(6):1934).
• Lin additionally appears on the 1990-filed Insystems alignment/HOE cases, and is the "L. H. Lin" of Collier, Burckhardt & Lin, Optical Holography (Academic Press, 1971) — i.e., a long-tenured optics principal, not a transient assignor.

Net: all three inventors remained with the original assignee for at least ~3.5 years post-filing, and two of them signed later Insystems patent applications. This is the opposite of the fire-sale precursor pattern.

---

Original assignee

Insystems, Inc. — 1120 Ringwood Court, San Jose, California 95131 (address per EP 0 294 643 A3 cover page and company-profile data).

• Primary line of business: developer/manufacturer of automated patterned-wafer defect-detection systems using laser holography and spatial frequency filtering — i.e., the commercial embodiment of the '774 disclosure. Founded 1985. Total disclosed raise ≈$4.6M; investors included Needham Capital Partners, Bayview and JPMorgan Fleming (CB Insights / Lighthouse company records). Field of use corroborated by Google Patents' own "Definitions" for this patent and by the Dataquest market-share tables in the Computer History Museum archive.
• Did it ship a product embodying the claims? Yes. The Insystems Model 8600 is described in the 1988 JVST paper as "currently in use for full wafer inspection at a number of semiconductor fabrication facilities in the U.S. and overseas," with a stated performance envelope of 0.5-µm defect sensitivity over a 150-mm patterned wafer in ~30 min — matching the '774 specification's Fourier-transform-lens + spatial-filter + inverse-transform + scanned CCD architecture. The later Model 8800 is described in Handbook of Optics (Vol. 1, §33.6) as the production implementation of the holographic filtering technique. Dataquest (May 1993) tables show Insystems carrying recognizable wafer-inspection revenue through roughly 1991 and zero by 1992 — a company that shipped real product and then ceased to be a going concern independently.
• Current status: No longer independent. PatSnap's company record states InSystems, Inc. is "Part of Optical Specialties, Inc." — consistent with the recorded 1993 transfer below. I found no bankruptcy filing for Insystems and, separately, no evidence that Optical Specialties, Inc. itself survives in its original form. One unverified lead: an SEC annual report (indexed via Onto Innovation's filing library) refers to a January 1998 acquisition "from Optical Specialties, Inc." — I could not confirm which acquirer or which assets without pulling the filing, so treat that as a lead, not a finding.

---

Assignment timeline

Methodology caveat first. I could not open assignmentcenter.uspto.gov / assignment.uspto.gov directly in this session, and my web queries did not return the reel/frame numbers or the correspondent of record for any of these conveyances. The chain below is taken from the Google Patents "Legal Events" mirror of the USPTO assignment record (https://patents.google.com/patent/US4806774/en), which reproduces the same conveyances and dates but not the reel/frame or correspondent fields. Reel/frame and correspondent are therefore recorded as "not retrieved" rather than invented. Note that the clause "if the Assignment Center has no records, say so and stop" does not apply — records do exist; only the metadata layer is missing.

1. 1988-04-12 (date as listed by the source; the mirror does not separate execution from recordation) — Reel not retrieved

   • Conveyance: Assignment of assignors' interest (originally recorded in Google as "ASSIGNMENT OF ASSIGNORS INTEREST")
   • Assignor: Cavan, Daniel L.; Howe, Robert B.; Lin, Lawrence H.
   • Assignee: Insystems, Inc.
   • Correspondent: not retrieved. (Distinct data point, not a substitute: Justia's "Patents Represented by Law Firm Stoel Rives Boley" page lists this patent — a single appearance, which under your own recurrence rule is not a finding.)
   • Context: ordinary inventor-to-company assignment, recorded ~10 months after the 1987-06-08 filing.

2. 1992-11-19 — Reel not retrieved

   • Conveyance: Security Interest ("see document for details")
   • Assignor: Optical Specialties, Inc.
   • Assignee: Silicon Valley Bank
   • Correspondent: not retrieved.
   • Context: securitization — a financing lien on the IP collateral of Optical Specialties, not a transfer of title. Note the sequencing: this lien is granted by Optical Specialties before the Insystems→OSI assignment records in 1993, which is consistent with OSI already controlling Insystems' IP at that date.

3. 1993-11-23 — Reel not retrieved

   • Conveyance: Assignment of interest ("see document for details")
   • Assignor: Insystems
   • Assignee: Optical Specialities, Inc. (spelled as recorded in the source; I am not normalizing it to "Specialties")
   • Correspondent: not retrieved.
   • Context: corporate absorption / internal consolidation — Insystems' patent estate folded into Optical Specialties, the then-operating optical CD/wafer-inspection vendor (OSI shipped the Metra 2000 overlay, Vista 1000 inspection and Inspectra 5000 defect systems per Dataquest, July 1989; it also ran a 1989 R&D/licensing arrangement with PruTech R&D Partnership II and Toray Industries).

4. 2007-06-08 — not an assignment; legal event only. Anticipated expiration for failure to maintain; status Ceased. Google Patents lists the current assignee as "OPTICAL SPECIALITIES Inc" — i.e., the terminal recorded owner is OSI, with no subsequent recorded transfer to any licensing entity.

Related but not an assignment: the reissue overlay — reissue application US 07/613,208 (priority 1990-11-14), granted 1992-06-09 as US RE33956 (further publication RE33956 F1, 1994-06-07). Reissue does not move title, but it means the 1989-issued claims were surrendered under 35 U.S.C. § 252 and replaced; any later chain of title attaches to the reissued claims.

Foreign-address data point (not a USPTO correspondent): the EP counterpart names the representative as Meddle, Alan Leonard et al, FORRESTER & BOEHMERT, Widemayerstrasse 4/I, D-8000 München 22 (DE). That is the EPO agent of record for Insystems, not an assignment-record correspondent, and it does not recur in the US record.

---

Timeline diagram

timeline
    title Ownership of US 4806774
    1987 : Application filed by Lin Cavan Howe
    1988 : Inventors assign to Insystems Inc
    1989 : Patent issued to Insystems Inc
    1992 : Security interest to Silicon Valley Bank
         : Reissue RE33956 granted
    1993 : Insystems IP assigned to Optical Specialties
    2007 : Patent expires and status Ceased

---

NPE / troll-pattern signals

Citation limitation, stated once and applying to all eight entries: reel/frame and correspondent fields were not retrievable in this session. Each call below is therefore cited to the conveyance type, parties and date as recorded in the Google Patents legal-events mirror of the USPTO assignment record, and to third-party corporate records. Where the missing reel/frame would be the dispositive evidence (signals 3 and 6), the call is unclear, not "not present."

1. Shell-entity transfer — not present. Every transfer node is an operating company or a bank: Insystems, Inc. → Optical Specialities, Inc. (1993-11-23), plus a lien to Silicon Valley Bank (1992-11-19). No assignee bears an "IP / Patents / Licensing / Holdings / Ventures" suffix, and Optical Specialties was demonstrably a product company (Metra 2000, Vista 1000, Inspectra 5000; Dataquest revenue tables 1984–1992).

2. Known asserter in the chain — not present. Neither Insystems, Inc. nor Optical Specialities, Inc. nor Silicon Valley Bank matches any entry on the reference list you supplied (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, Spangenberg entities) or any entity surfaced in my searches as a high-frequency plaintiff. The SVB entry is a collateral lien by a commercial bank, which is the standard financing signature of an operating company, not an asserter.

3. Repeat correspondent across the chain — unclear (cannot be tested). With only two substantive conveyances sharing one corporate family and a security interest, the recurrence test is weak even if the data were complete. The single attributable name in the record is Stoel Rives Boley (Justia's prosecution-firm attribution for this patent) — one appearance, which by your explicit rule is not a finding. No correspondent name is available for any of the three recorded conveyances. No finding.

4. Cascading transfers — not present. Two recorded conveyances in the ~6 years between issuance (1989-02-21) and the 1993 assignment; the entire chain spans 1988 → 1993, not "<24 months through chained LLCs." No shared-correspondent-address pattern is visible because no addresses are visible, but the number of transfers alone rules out a cascade.

5. Pre-litigation transfer — not present. No infringement suit naming this patent was identified in the prior litigation workup, so there is no "first suit" date for an assignment to precede. Independently, the terminal assignment (1993-11-23) predates expiry (2007-06-08) by ~14 years; a pre-litigation venue/standing transfer is not the plausible reading of this chain.

6. Bankruptcy fire-sale — unclear. The circumstantial case is real: Insystems' wafer-inspection revenue collapses to zero by 1992 in the Dataquest tables, and the 1992 SVB lien was granted by Optical Specialties on assets that Insystems still nominally held until the 1993 assignment — a signature of distress or of a lender already looking through to the operating parent. But I found no Chapter 7/11 filing, no § 363 sale order, and no assignment recital of bankruptcy for either Insystems or Optical Specialties. Distress is supported; a bankruptcy sale is not. Marked unclear rather than present.

7. Privateering — not present. Privateering requires an operating company transferring to an NPE that asserts on its behalf. Here the operating company (Insystems) transferred to another operating company (Optical Specialties), which was itself a competitor in the same wafer-inspection market per the Dataquest rankings. No NPE, no assertion, no SEC/EFF/Patent Progress coverage surfaced.

8. Defensive aggregator (anti-NPE) — not present. The chain terminates at Optical Specialities, Inc., not at RPX, AST, LOT Network, Unified Patents or OIN. No defensive aggregator appears anywhere in the ownership record or in the successor-entity data I was able to retrieve. (The patent is cited as prior art by later third parties, e.g. in the ISR for WO 2013/114022 and EP 1 231 460 A3, and is the acknowledged background in US 5,172,000 — that is public-domain citation, not defensive acquisition.)

---

Verdict

Insufficient data.

Justification: the recorded chain consists of exactly two substantive events — a 1992-11-19 security interest to Silicon Valley Bank and a 1993-11-23 assignment from Insystems to Optical Specialities, Inc. — both between operating companies in the wafer-inspection business, with zero of the eight NPE signals present and none of the classic asserter markers (no shell LLC, no licensing-only assignee, no aggregator, no privateering, no litigation). That pattern is affirmatively anti-NPE, and it is why the answer is not "NPE — moderate confidence" despite the 1992–93 distress. The verdict is nevertheless Insufficient data rather than a clean negative because (a) reel/frame and correspondent of record were not retrievable, so signals 3 (repeat correspondent) and 6 (bankruptcy fire-sale) could not actually be tested, and (b) the terminal owner's modern identity is unverified — Insystems is documented only as "part of Optical Specialties, Inc.," and Optical Specialities, Inc. itself has no confirmed 2026 existence, leaving open whether an unrecorded transfer ever occurred. Note the built-in safety valve: any post-1993 transfer would have had to be recorded to be effective against third parties, and the Assignment Center record — mirrored on Google Patents — shows none; combined with the 2007-06-08 expiry and "Ceased" status, this patent could not support an NPE campaign in 2026 regardless of how the last twenty years of title are read.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (legacy interface: https://assignment.uspto.gov/patent/index.html), search patent number 4806774; cross-check the reissue RE33956 and the '208 reissue application, and pull the maintenance-fee and "Ceased" status from USPTO Patent Center (https://patentcenter.uspto.gov/).

I'll research the prior-art citations of record for US 4,806,774, including both the references it cites (backward) and the published front-page references.

Let me get the actual front-page "References Cited" for US 4,806,774 and the full citations list.

Let me pull the actual front-page references cited for US 4,806,774 and verify the EP search-report citations.

Prior Art Analysis — U.S. Patent No. 4,806,774

Status of the underlying record — read this first. Two data gaps materially limit this analysis, and I am not going to paper over them:

1. The literal claim set of U.S. 4,806,774 was not retrievable. Consistent with the earlier sections of this analysis, the supplied patent text breaks off mid-Detailed Description (in the Time-Delay-Integration passage, before the claims). Every "which claim does it anticipate" statement below is therefore keyed to claim concepts reconstructed from the specification, the abstract, and the EP counterpart — not to literal claim language. Do not quote them as claim text.
2. I could not confirm the complete front-page "(56) References Cited" list for 4,806,774 itself. My searches returned the citation table for US 5,428,442 (a later Insystems patent that cites '774), not '774's own face. I therefore separate below (a) references the '774 specification expressly discusses, which is confirmed, from (b) references appearing in the EP 0 294 643 search report, carried over from the earlier-generated section, and (c) field references whose status as cited art against '774 I could not verify.

I did not substitute any similar patent number, and I have not auto-corrected any identifier.

---

1. Bibliographic confirmation (USPTO/Google Patents)

Because the '774 patent was filed on June 8, 1987, it is a pre-AIA case: prior-art analysis runs under 35 U.S.C. §§ 102(a), 102(b), 102(e) and 103 as they stood before the AIA, and the reissue overlay means the operative (reissue) claims are what a § 102 comparison should ultimately be run against.

---

2. The references that matter

A. References expressly identified in the '774 specification (confirmed)

These are the two references the patent itself frames as the closest art, in its BACKGROUND OF THE INVENTION. Both are also of-record citations in the U.S. file and appear on the face of the patent.

A-1. U.S. Patent No. 4,000,949 — Watkins

A-2. U.S. Patent No. 3,614,232 — Mathisen

Why the '774 patent exists at all: the specification expressly states that Watkins and Mathisen "imply that systems of the type they describe require both inter- and intra-die interference pattern information," and that the inter-die component consists of "very closely spaced light spots that are extremely difficult to resolve by a Fourier transform lens." The '774 invention's asserted departure is the deliberate use of intra-die information only, restricted to dies having many redundant circuit patterns, which permits a small-object-field lens and a near-diffraction-limited transform. That is precisely the line of demarcation against A-1 and A-2.

---

B. Non-patent literature cited (confirmed in the record)

B-1. Watkins, "Inspection of Integrated Circuit Photomasks with Intensity Spatial Filters"

B-2. Kahn, "Large Area, Engineering Drawing Quality Displays Using Laser Addressed Smectic Liquid Crystal Light Valves"

---

C. References appearing in the EP 0 294 643 search report (carried forward from the earlier section — not re-verified in this session)

The European search report is reported to have cited claims 1–4, 6, 7, 10–20, 24 and 28 against art including:

I am flagging these three explicitly as unverified here. The earlier-generated section states they were cited, and I am building on that rather than repeating it, but I could not independently confirm them before exhausting the search budget. They should be pulled from the EP 0 294 643 B1 search-report annex before being relied on.

---

D. Field references — I could NOT confirm these were cited against '774

The searches surfaced these as prominent background art in the same art unit, commonly cited by other patents in the spatial-filtering family (e.g., they are discussed in US 4,516,833 and in the reviews of the field). I found no confirmation that any of them appears on the face of U.S. 4,806,774, and I am not asserting that they do:

• U.S. 3,435,244 — Burckhardt et al. (cited within Watkins '949).
• U.S. 3,658,420 — Axelrod (1972); increased-opaque-area spatial filters.
• U.S. 3,738,752 — Heinz et al. (1973); computer-generated filters, deviation of image points from the linear relation.
• U.S. 3,790,280 — Heinz et al. (1974); spatial filter with non-uniform region spacing, compensating regions.
• U.S. 3,972,616 — Minami et al. (1976); wavelength-selective spatial filter, movable mask driver.
• U.S. 4,516,833 / EP 0 114 524 A1 — Manufacture of high performance optical spatial filter (multi-generation photographic filters; method of isolating non-periodic defects). This is a later-to-'774-family filter-fabrication patent that itself cites Watkins '949 and Mathisen '232; it is useful as § 103 art against filter-fabrication dependent claims only if it predates the relevant '774 claim (its EP publication, 1984-08-01, does predate the 1987 filing).
• Also of-record-adjacent: US 4,197,011 (RCA, Defect detection and plotting system, 1980-04-08) and US 4,636,069 (Matrix Instruments, 1987-01-13) — both appeared in a citation table on the US 5,428,442 page (which cites '774), not confirmed as '774's own citations.

Primary NPL for this field, useful if you need to reconstruct the state of the art: L. S. Watkins, Solid State Technology 12(2):35–38 (1969) and 12(9):29–32 (1969); R. A. Heinz, R. L. Odenweller, R. C. Oehrle and L. S. Watkins, Western Electric Engineer 17:37–56 (1973); N. N. Axelrod, Proc. IEEE 60:447–448 (1972).

---

3. Anticipation matrix (against reconstructed claim concepts)

No single reference above discloses all elements of the reconstructed independent claims. The distinguishing combination — inverse-transform lens forming a magnified defect image onto a 2-D photodetector array deliberately sized and centred so that only the on-axis die's defects are captured, with continuous X-Y stage raster scanning under time-delay integration — is not met by Watkins '949 or Mathisen '232. That is exactly what the specification says: the '774 patent's whole point is discarding the inter-die information its predecessors needed.

---

4. § 102 vs. § 103 — practical read

• § 102 (anticipation): only realistically available against the broadest independent claims, and only if those claims are drafted without the stage/TDI/limited-detector limitations. On the reconstructed claim 1 (method) and claim 19 (system), Watkins '949 and Mathisen '232 are single-reference anticipation candidates only in theory; a well-drafted claim would survive them.
• § 103 (obviousness): the realistic attack. Watkins '949 or Mathisen '232 as the primary reference, combined with a scanning-stage/TDI reference (the '774 spec itself names Kensington Laboratories Model 8500 X-Y table and the RCA Model 6220-004 CCD (403 × 512)), would be the natural combination. Note, however, that this is a pre-AIA case, so the motivation-to-combine analysis is the older, more permissive standard — and the patent owner's counter is the specification's own express teaching away from inter-die information, which cuts against combining the array-inspection references with a die-scanning architecture.
• Important overlay: because the '774 claims were surrendered and replaced by the 1992 reissue (RE33,956) under 35 U.S.C. § 252, any serious § 102 analysis must ultimately be run against the reissue claim set, which may differ from the original — and I could not retrieve either claim set literally.

---

5. Explicit uncertainties in this section

1. The literal claims of US 4,806,774 are still not confirmed. All § 102 statements are keyed to reconstructed claim concepts and must be re-run against the granted/reissue claim text.
2. The complete (56) reference list on the '774 face was not retrievable. What I have empirically confirmed is: (a) the two references the specification discusses (Watkins '949, Mathisen '232); (b) the Watkins Proc. IEEE NPL item; (c) the Kahn light-valve citation. The front-page patent-citation table (patent numbers, dates, classes) is inferred/absent.
3. The EP search-report citations (EP 0 017 371, US 4,370,024, Sterling) are carried forward from the earlier section and were not re-verified here. Verify against the EP 0 294 643 B1 search-report annex.
4. US 4,197,011 and US 4,636,069 appeared in a citation table belonging to US 5,428,442, not confirmed as '774's own; I am not asserting them as of-record citations of '774.
5. Assignee spelling for the current owner remains as given in the record — "Optical Specialities, Inc." — and I am not normalizing it.
6. No litigation or PTAB record exists for this patent to inform a secondary-considerations analysis; as noted in the earlier sections, the patent is Ceased with anticipated expiration 2007-06-08 and was surrendered-by-reissue in 1992.

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

Obviousness Analysis Under 35 U.S.C. § 103 — U.S. Patent No. 4,806,774

Patent analyzed: U.S. 4,806,774 A ("the '774 patent"), Inspection system for array of microcircuit dies having redundant circuit patterns, app. 07/060,090, filed 1987‑06‑08, granted 1989‑02‑21, inventors Lin/Cavan/Howe, original assignee Insystems, Inc.
Source of record: https://patents.google.com/patent/[US4806774](/patent/US4806774)/en

---

1. Threshold caveats and cross-references

Claim text unavailable — analysis is architecture-based. As flagged in the previously generated Patent Summary, the supplied '774 record breaks off mid‑Detailed Description (in the TDI section), before the claims. The EP counterpart (EP 0 294 643 B1) states the invention provides "a detection method having the features of claim 1" and "a detection system having the features of claim 19," and the EP search report cited claims 1–4, 6, 7, 10–20, 24 and 28. Everything below is therefore keyed to the reconstructed independent claim architecture (method + system) from that earlier section, not to verbatim claim language. This is the single largest limitation on the reliability of this analysis and should be cured against the granted PDF and against RE33956 before any conclusion is treated as final.

The claim set that matters may be the reissue's, not the '774's. Under 35 U.S.C. § 252, the original '774 claims were surrendered upon reissue (RE33956, granted 1992‑06‑09). An obviousness attack on "the '774 claims" in 2026 is partly academic; the earlier PTAB/Strategic sections correctly identify the reissue as the operative narrowing event. That said, the substantive § 103 question — was the Lin/Cavan/Howe contribution obvious over Watkins/Mathisen/Tokyo‑Shibaura/Task — is the same question for both original and reissue claims.

Contradiction flagged. The task prompt states the current date is April 26, 2026; the previously generated sections and the patent fetch are dated 2026‑09‑30. Immaterial to the § 103 analysis, but noted per instruction.

Also flagged: the Google Patents "Prior art date" field reads 1987‑06‑08 — that is the filing date, not a separate earlier priority date. The '774 patent is pre‑AIA; obviousness is governed by pre‑AIA § 103, though modern courts apply KSR's flexible framework to pre‑AIA patents as well.

---

2. The legal framework applied

Level of ordinary skill (PHOSITA), as of June 1987: a B.S./M.S. in optics or electrical engineering with 2–3 years' experience in coherent-optical inspection systems, or equivalent; familiar with Fourier-plane spatial filtering, photographic and liquid-crystal spatial filters, lens aberration theory (isoplanatism, telecentricity), and area-array CCD imaging with TDI. The '774 specification itself presumes exactly this person.

Caveat: I could not retrieve the actual examiner's rejections or applicant's remarks from the '774 file wrapper. The grounds below are constructed from the references; they are not a reconstruction of what the examiner did or said.

---

3. The references (prior-art record, with sources)

Two admissions inside the '774 specification itself are also prior art against its own claims: spatial filters may be made by "known computer generation techniques"; the reflective filter is a laser smectic light valve per Kahn, "Large Area, Engineering Drawing Quality Displays Using Laser Addressed Smectic Liquid Crystal Light Valves," Automation Technology Institute Conference, Montreal, February 1987; and the X‑Y table (Kensington Laboratories Model 8500) and CCD (RCA Model 6220‑004) are named commercial parts.

---

4. Reconstructed claim elements vs. the art

On element H — the heart of the case — three points matter.

1. The intra-die approach is disclosed, not merely suggested. Watkins' FIG. 7 filter is built from the diffraction pattern of a single element (i.e., intra-die information) and is expressly described as an operative alternative. A claim to the alternative is obvious over the reference that discloses it as a workable option (MPEP § 2131).
2. The illumination/field-size tradeoff is quantitatively taught. Iwamoto gives the FT spot geometry 2λf/NP and λf/P(1−2/N). A PHOSITA reading that would immediately see that the FT spots become smaller and better separated as the number of illuminated periods N rises. '774's own design rationale — illuminate a 20 mm area "to develop an accurate Fourier transform pattern of the redundant circuit patterns" while using a lens of only 3 mm object field — is the direct, predictable exploitation of that known relation. The '774 resolution criterion d₁ ≪ λf₁/c is the same classical diffraction-limit relation.
3. The prior art already treated the alignment burden as the problem to be solved. EP 0 017 371's background enumerates the three intolerable requirements of precise-matched-filter systems (orientation alignment, high-precision low-distortion FT lens, precise axial positioning). '774's premise — "inter-die interference pattern information is unnecessary if only areas of many redundant patterns are inspected" — is the natural resolution of exactly that stated problem.

---

5. Grounds of rejection

Ground 1 — Watkins in view of Mathisen (primary combination; § 103(a))

Watkins supplies A, B, C, D and the express teaching that the composite diffraction pattern is separable into inter-die and intra-die components, plus two alternative filter constructions (spacing-only; single-element). Mathisen supplies the imaging/detection and automation layer: an imaging detector (vidicon) with a critical-area mask that limits detection to selected regions of the image field (element E), automatic centering (addressing the alignment problem), and scan generation.

Motivation. Same field, same problem, same physics; both are directed to detecting nonperiodic defects in repetitive microcircuit patterns by Fourier-plane filtering. KSR holds combination obvious where each element does no more than its established function: the FT lens transforms; the matched filter blocks the periodic spectrum; the re-imaging lens restores real space; the masked detector reads out the on-axis region. Nothing new is asked of any element.

Reasonable expectation of success. Watkins expressly teaches that the FIG. 7 (single-element) filter may be made by exposing a plate at the focal plane with a good mask or by theoretical calculation of the diffraction pattern of a perfect mask — so the filter for the intra-die spectrum is known to be constructible, and by a differently ordered arrangement if only one die's field is imaged.

Ground 2 — Watkins and/or Mathisen in view of Tokyo Shibaura (EP 0 017 371 / US 4,330,775 / US 4,360,269) (the "intra-die" combination)

This is the ground that most directly reaches element H. Iwamoto discloses the complete sequence A→B→C→D→E for periodic patterns including semiconductor products, with a filter that by design rejects the pattern's own periodic spectra and transmits defect energy — functional equivalence to "intra-die interference pattern information only." It adds a reflective-mode beam-splitter variant, a photoelectric converter with pinhole screen at the image plane, and specimen rotation/translation scanning.

Motivation. (i) Iwamoto's express problem statement is the alignment/resolution burden of the precise-matched-filter art; using only the local (intra-die) periodic structure with an alignment-insensitive filter is the identified solution. (ii) Iwamoto's diffraction formulas make the N-period resolution tradeoff explicit, supplying the design rule for illuminating many redundant periods while restricting the imaged field. (iii) The '774 patent itself, in its EP counterpart's background, characterizes the Iwamoto system as a known spatial-filter inspection system — an acknowledgment that it is analogous art in the same field. (iv) KSR "obvious to try": a PHOSITA faced with the inter-die spot-resolution problem has a finite number of identified, predictable approaches — filter the inter-die interference function (Watkins FIG. 6), filter the composite (Watkins FIG. 7 / Mathisen), or filter the intra-pattern periodic spectrum of an illuminated patch (Iwamoto).

Ground 3 — Any of Grounds 1–2 in view of Task (US 4,370,024) (the reflective liquid-crystal embodiment)

'774's second embodiment (FIGS. 6–7) uses a reflective liquid-crystal (laser smectic light valve) spatial filter 102 in a folded configuration with a polarizing beam splitter and two quarter-wave plates. Task discloses the same physical scheme at the same point in the system: a coherent beam, a Fourier-transform lens system, a beam splitter folding the axis onto the reflective face of a liquid-crystal light valve situated in the Fourier plane, a programmable write pattern, and an inverse-transform lens system to a real-space image plane — with a claim expressly requiring a reflective region at the zero-spatial-frequency point.

Motivation. Task's programmable filter solves the practical objection to Watkins' photographic filter (one plate per product; fabrication burden; the intensity-range problem identified in the JPS59170876A/EP 0 011 524 literature). A PHOSITA wanting a reusable filter in an inspection tool would adopt the reflective LC valve configuration, fold the path with a beam splitter, and (to keep the reflected defect beam separated from the illumination and to recover the polarization state) add the quarter-wave/polarizing-beam-splitter arrangement — the standard, and only, way to make a reflection-mode polarized Fourier filter work. The specific LC cell construction (Kahn's laser-addressed smectic valve) is admitted prior art by the '774 specification itself.

Field-of-endeavor objection anticipated: Task is arguably not wafer-inspection art. Under KSR that is not dispositive — the reference is reasonably pertinent because it addresses the same problem (dynamically filtering a coherent beam in its Fourier plane) with the same structure, and is therefore combinable.

Ground 4 — Grounds 1–3 further in view of the secondary spatial-filter literature (Heinz et al.)

'774's central engineering claims are (a) a near-diffraction-limited, isoplanatic/aplanatic/anastigmatic Fourier-transform lens over ±15°–20°, and (b) a lens design that anticipates and cancels the transform lens' residual aberrations in the inverse transform. Heinz (US 3,738,752; US 3,790,280) already recognized that real lenses displace FT spots from the ideal x = nλf/d positions and that "aberrations and distortions that exist in actual optical systems must be taken into account when creating a filter," including coma, astigmatism, field curvature and distortion. That is the identical problem '774 solves by lens design rather than filter design. Once the problem is known and stated in these terms, telecentricity, isoplanatism and aberration-balancing between two lens groups are textbook design objectives.

Ground 5 — Grounds 1–4 in view of conventional area-array CCD practice (elements F and G)

• Translation stage (F): Iwamoto claims shifting and rotating means; Watkins suggests rectilinear displacement "for scanning purposes"; Minami uses a mask driver. A precision X‑Y stage is named as an off-the-shelf commercial item (Kensington 8500).
• On-axis detector smaller than the image field (E): functionally identical to Mathisen's critical-area mask; alternatively, simply selecting a commercially available CCD (RCA 6220‑004) and choosing magnification to match pixel size to the resolution limit — which '774 itself says is the design rule ("magnification… of an amount that approximately matches the resolution limit of the image to the pixel size").
• TDI raster scanning (G): the weakest-disclosed element. But KSR's "known technique to improve similar devices in the same way" applies squarely: TDI is the standard way to raise signal-to-noise while scanning continuously, and the benefit of continuous (versus step-and-repeat) stage motion — throughput — is a routine engineering tradeoff. That the patent uses a commercially catalogued CCD and a commercially catalogued X‑Y table is In re Harza-style selection of known components.

Ground 6 — The numerical prescriptions (Tables I–IV) are not separately patentable

If any claim recites the specific prescriptions (f₁ = 60 mm, f₂ = 600 mm, m = 10, 442.5 nm, 20 mm illumination, 3 mm object field, ±15°–20° telecentric cone, 403 × 512 elements at 16 × 20 µm), these are results of routine optimization against stated performance requirements. In re Aller and In re Wood: discovering optimum or workable values of known parameters by routine experimentation is not invention. The specification itself concedes the endpoint is compromised — "the image quality at image plane 60 marginally meets the design objectives because of the limitations imposed by such imbalance of field curvature and astigmatism and the existence of higher order residual spherical aberrations." That concession is doubly damaging: it defeats "unexpected results" and undermines any argument that the design constitutes an unforeseeable achievement.

---

6. Anticipation-adjacent observations (§ 102)

If the reconstructed claim 1 is drafted without the element-H "intra-die only / on-axis die" restriction, then Watkins alone or Tokyo Shibaura alone may anticipate: both disclose coherent illumination of a patterned specimen, a Fourier-transform lens, a pre-fabricated filter at the transform plane blocking the error-free periodic spectrum, an inverse-transform/re-imaging lens, and detection of the resulting defect image. I flag this because the earlier section notes the 1992 reissue (RE33956) — under § 251(a) a reissue can only narrow, which suggests the originally issued claims were believed to be over-broad relative to this art. If so, the original '774 claims were likely anticipated or plainly obvious, and the reissue claims (which I have not seen) are the only ones worth attacking. I cannot verify the reissue claim content and am not asserting it as fact.

---

7. Teaching away, and the strongest patent-owner positions

There is no true teaching away. Watkins and Mathisen disclose that the composite diffraction pattern contains both components; they do not disparage, criticize, or discourage the use of intra-die information alone, and Watkins expressly offers a single-element-based filter as a usable alternative. Under In re Fulton and Medichem, disclosure of a preferred approach is not a teaching away. Iwamoto's stated object — reducing the alignment burden of precisely matched filters — affirmatively points toward the '774 approach.

The realistic patent-owner positions are:

1. Recognition of the problem ("the inter-die information is unnecessary"). '774's contribution is largely the insight that one may discard the inter-die interference function when inspecting dies containing many redundant patterns. That recognition is the patent's best story. It is a weak legal position under KSR ("the question is not whether the combination was obvious to the patentee") — but it is the honest one, and it is the argument the specification advances.
2. The specific combination of a large illumination aperture with a small-object-field lens. No single reference of record appears to state this exact tradeoff. But Iwamoto's N-dependence and the '774 specification's own admission that the tradeoff was made "to minimize the introduction of aberrations" reduce this to an optimization.
3. The TDI + stripe raster implementation. The most concrete undisclosed detail; but correspondingly the most clearly a routine engineering choice.
4. Prosecution history as objective evidence of nonobviousness. The examiner allowed the case in 1988–89 and EP 0 294 643 B1 granted, with the EP search report having cited claims 1–4, 6, 7, 10–20, 24 and 28 against prior art. That is real, but it is weak evidence: I cannot tell from the record whether Watkins, Mathisen, the Tokyo Shibaura family, Task, and the Sterling XDJ item were all made of record and considered, and a granted EP case is not an adjudication of § 103 with the benefit of KSR.

No objective evidence of nonobviousness is established on this record. There is no verified long-felt-need, commercial-success, copying, or industry-praise evidence with a nexus to the claimed subject matter. The corporate history (Insystems → security interest to Silicon Valley Bank 1992 → assignment to Optical Specialties 1993 → "Ceased," expired 2007‑06‑08) is consistent with a product business, not with a patent whose validity was tested and vindicated. Conversely, the patent's own "marginally meets the design objectives" statement affirmatively cuts against unexpected results.

---

8. Conclusion

On the record available, and on the reconstructed claim architecture, the '774 claims would very likely have been held obvious under § 103. The strongest ground is Watkins + Mathisen + Tokyo Shibaura (EP 0 017 371 / US 4,330,775 / US 4,360,269), optionally further in view of Task (US 4,370,024) for the reflective liquid-crystal filter embodiment, with routine CCD/TDI and X‑Y-stage practice supplying the scanning/detection elements. The reasons are the classic ones: a single field of endeavor, a shared and expressly stated problem (filter alignment and FT spot resolution), a finite number of identified and predictable solutions, express disclosure of the intra-die alternative in the primary reference, quantitative teaching in Iwamoto of the illumination-N versus FT-spot-resolution relationship that underlies '774's core design choice, and a specification that admits its own endpoint is only marginally achieved.

The best, and probably only, serious nonobviousness argument is claim-scope-dependent: if the reissue claims (RE33956) recite the specific
large-aperture-illumination / small-object-field lens architecture, or the specific TDI stripe-scan/on-axis-CCD combination, in a way that no reference or combination reaches, the case shifts from "plainly obvious" to "close." That cannot be assessed without the actual claim text.

Explicit uncertainties carried forward:

1. Claim language is unverified — the single biggest limitation; all element designations are reconstructions.
2. The reissue claims (RE33956) were not examined; they, not the original '774 claims, are the operative narrowing.
3. The Xerox Disclosure Journal item (Sterling et al.) was not retrievable and is relied on nowhere above.
4. The '774 file wrapper (examiner's rejections and applicant's remarks) was not retrieved, so this is not a reconstruction of the actual prosecution.
5. Iwamoto family identification is inferential in part: I verified US 4,330,775 and US 4,360,269 as Tokyo Shibaura/Iwamoto periodic-pattern spatial-filter systems, and the EP 0 294 643 German text (DE 3888395 T2) attributes a spatial-filter inspection system to "Iwamoto" in the European application numbered "0 017 317" (which I read as a typo for 0 017 371 — I am not auto-correcting it, and I have not confirmed the number as printed in the original).
6. Date discrepancy between the task prompt (2026‑04‑26) and the source/prior sections (2026‑09‑30), noted and immaterial.