US 7092573

I'll search for this specific patent number and check for any CAFC 2026 docket activity.

The first search returned good results. Let me dig deeper on the specific litigation and any 2026 appellate activity.

Let me verify whether there's any 2026 Federal Circuit activity specifically involving this patent.

US Patent 7,092,573 — Summary

Bibliographic Data (per the authoritative full text and USPTO/Google Patents record)

Abstract

"A method for processing a digital color image includes the steps of: providing a subject matter detector for distinguishing between target and background subject matters; applying the subject matter detector to the image to produce a belief map indicating the degree of belief that pixels in the image belong to target subject matter; providing an image enhancement operation that is responsive to a control signal for controlling the degree of image enhancement; and applying image enhancement to the digital image by varying the control signal according to the belief map to produce an enhanced image."

Independent Claims — Plain-Language Overview

The patent has five claims in independent form plus one product claim (claims 1, 7, 8, 9, 13, and 32). All share the core "belief map → control signal → spatially varying enhancement" architecture; they differ in what belief-map features feed the control signal.

• Claim 1 (method). Apply a subject-matter detector to the image to produce a belief map of values indicating the degree of belief that pixels belong to target subject matter; the values define "belief regions." Determine the size of each belief region, then enhance the image, with the enhancement varying pixel by pixel in accordance with both the degree of belief and the size of the respective belief region.

• Claim 7 (system). Apparatus counterpart of claim 1: a subject-matter detector that produces the belief map; a belief map analyzer that determines the sizes of the belief regions; and an image-enhancement operator that controls the degree of enhancement pixel by pixel according to both the degree of belief and the region size.

• Claim 8 (method). Like claim 1, but the method determines the location and size of each belief region, and the pixel-by-pixel enhancement varies according to the degree of belief, size, and location of the respective belief region.

• Claim 9 (method). Provide a subject-matter detector distinguishing target vs. background; apply it to produce a belief map; analyze the belief map to generate a control signal; provide an enhancement operation responsive to that control signal; apply the enhancement while varying the control signal pixel by pixel according to the belief map — wherein the analysis includes determining the location and size of each belief region and enhancing the control signal based on the locations.

• Claim 13 (method). Same overall pipeline as claim 9, but the wherein clause recites that the analysis includes determining the size of each belief region and enhancing the control signal based on the size. This is the claim from which most of the specific subject-matter and enhancement-type limitations depend (see below), and it is the one whose "size" requirement has drawn the most litigation attention.

• Claim 32 (product). "A computer readable program product stored on a computer readable medium for performing the method of claim 13."

Notable dependent claims (context for scope)

• Claim 15: target subject matters include human flesh, clear blue sky, lawn grass, snow fields, and water bodies.
• Claim 17: enhancement operations include sharpening, noise reduction, JPEG de-blocking, tone scale adjustment, scene balance adjustment, and color re-mapping.
• Claims 23–28: each recites a single enhancement type (sharpening; noise reduction; tone scale; scene balance; color re-mapping; JPEG de-blocking).
• Claims 29–30: image magnification via interpolation, selectable between bilinear and fractal-based interpolation.
• Claim 33: belief values indicate more than two different degrees of belief.
• Claims 4–6, 10–12, 14, 16, 31 cover selecting among multiple subject-matter detectors / enhancement operations and optional resolution reduction before detection.

Litigation and Post-Grant Activity (as reflected in the record)

• Monument Peak Ventures, LLC v. GE Healthcare, Inc., No. 3:18-cv-01158 (S.D. Cal., filed June 4, 2018) — '573 asserted (along with 7,062,085; 7,212,668; 6,509,910; 8,984,419); claim-construction briefing filed May 2019; case stayed pending IPR.
• General Electric Co. et al. v. Monument Peak Ventures, LLC, IPR2019-01020 (PTAB, petition filed April 25, 2019) — challenged U.S. 7,092,573 (Luo et al.). Institution denied September 23, 2019 ("Decision Denying Institution of Inter Partes Review"). The Google Patents record likewise flags this PTAB case as "Not Instituted – Merits."
• MPV v. Xerox Corp., No. 2:21-cv-00345 (E.D. Tex., filed September 1, 2021) — '573 among six asserted patents.
• MPV v. TCL Electronics Holdings Ltd. et al., No. 5:24-cv-00011 (E.D. Tex., filed January 22, 2024) — '573 and 7,683,962 asserted against TCL smartphones; independent claim 1 pleaded against "beauty mode"/"AI scene detection"/"AI-powered facial detection" and region-size-based skin smoothing. A magistrate judge recommended denying TCL's motion to dismiss (service / failure to state a claim).
• Other district-court matters listed in the record include No. 2:23-cv-00169 and 2:23-cv-00127 (E.D. Tex.), and MPV v. Carl Zeiss AG et al., No. 2:26-cv-00100 (E.D. Tex., filed February 6, 2026).

CAFC 2026 Activity — Explicit Uncertainty

I searched specifically for Federal Circuit docket activity in 2026 tied to 7,092,573 (and for Monument Peak Ventures appeals generally). I found no authoritative evidence of a Court of Appeals for the Federal Circuit case in 2026 involving U.S. 7,092,573. Search hits returned unrelated 2026 CAFC matters (e.g., Ascendis v. BioMarin) and unrelated Texas patent appeals (e.g., Headwater Research v. Verizon), none of which concern the '573 patent. I cannot rule out a very recent or unindexed appeal, but I have no basis to state that one exists. This should be treated as an open item rather than a confirmed "no appeal" finding.

Caveats

• The Google Patents "family has litigation" and litigation entries are third-party aggregations (Unified Patents, Darts-ip) and may lag or contain errors; they are not a substitute for PACER/PTAB dockets.
• The Google Patents legal-status label ("Expired – Lifetime"; adjusted expiration 2023-07-13) is expressly flagged by Google as an assumption, not a legal conclusion. Given the December 10, 2001 filing date and the term extension, the patent has almost certainly expired, but confirm against the official USPTO Patent Term Adjustment/expiration record if the date matters for your purpose.
• Inventor name "Amit Singhal" appears on the patent and in the assignment record; it is distinct from any similarly named search-engine engineer and should not be conflated.

I'll research the litigation history for US 7,092,573. Let me start with the sources you mentioned and the patent's own litigation records.

Let me dig deeper into the specific docket numbers and case details.

Litigation and PTAB Review Involving U.S. Patent No. 7,092,573

Patent identification (as literally recorded): US 7,092,573 B2, "Method and system for selectively applying enhancement to an image," application Ser. No. 10/016,601, priority/filing date December 10, 2001, granted August 15, 2006. Original assignee Eastman Kodak Company; ownership passed to Intellectual Ventures Fund 83 LLC (2013) and then to Monument Peak Ventures, LLC (2017) — the current listed assignee, which is the plaintiff/patent owner in the district-court matters below. Listed status: Expired – Lifetime, adjusted expiration July 13, 2023.

The Google Patents record for this patent expressly lists litigation, and the Unified Patents PTAB entry (IPR2019-01020, "Not Instituted – Merits") is the only adversarial patent-office proceeding. Below is what I could confirm, with sources.

I. District Court Litigation

Related proceeding (not a '573 case): Xerox Corp. v. Monument Peak Ventures, LLC, No. 6:20-cv-6263 (W.D.N.Y.), a declaratory-judgment action transferred to the E.D. Tex. (Civil Action No. 4:21-cv-00683-SDJ) and voluntarily dismissed without prejudice on Sept. 1, 2021. The IPR petition documents show this DJ action concerned U.S. 7,684,090, not the '573 patent, so I list it only as context for the MPV–Xerox dispute.

II. PTAB (Inter Partes Review)

• The GE petition (Paper 1) challenged claims 1, 2, 4–6 and 8 as obvious primarily over U.S. 5,978,100 to Kinjo (called "Kinjo-569"), in various combinations with the Luo publications and U.S. 5,754,348 to Soohoo (cited as "Soohoo"). The Board declined to institute.
• This IPR ran in parallel with the S.D. Cal. GE litigation (#1 above), which was stayed pending the institution decision and then settled/dismissed with prejudice in October 2019.

III. What is NOT present

• No Court of Appeals for the Federal Circuit decision involving '573 appears in the records I reviewed. (An institution denial is generally non-appealable, and the district cases above were resolved by dismissal/settlement rather than judgment.)
• No ITC (Section 337) investigation involving '573 was found.
• The patent's EP counterpart is EP1318475B1 (and DE60203871T2); I found no separate EPO opposition outcome.

IV. Caveats on confidence

• Cases #1, #3, #4, #5, and the IPR are supported by multiple sources (Google Patents litigation links, Unified Patents, Docket Alarm, CourtListener, RPX, PatSnap, Ex Parte AI). Confidence: high.
• Case #2 (MPV v. Xerox, 2:21-cv-00345) is well-documented as filed with '573 asserted; its final outcome is not confirmed. Confidence in the filing: high; in the outcome: low.
• Case #6 (2:22-cv-00360) is listed in the patent's linked litigation data, but I could not confirm the defendant, and I will not guess. If you need this one pinned down, it should be pulled directly from PACER or the Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/caselist) filtered to the '573 family.
• All "dismissed with prejudice" outcomes reflect the docketed basis of termination; the underlying settlement terms are not public and I am not inferring any license or payment.
• The patent is shown as expired (adjusted expiration July 13, 2023), which affects the practical exposure of the 2023–2024 assertions but does not, by itself, resolve any of their dockets.

Primary sources used:

• Google Patents record for US 7,092,573 (litigation links): https://patents.google.com/patent/US7092573/en
• Unified Patents PTAB portal / IPR2019-01020: https://portal.unifiedpatents.com/ptab/case/IPR2019-01020
• IPR decision and documents: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2019-01020 and https://services.patexia.com/lawsuits/General-Electric-Co-et-al-v-Monument-Peak-Ventures-LLC-id-[124467](/patent/124467)/documents/
• S.D. Cal. docket (GE): https://www.courtlistener.com/docket/[7312391](/patent/7312391)/monument-peak-ventures-llc-v-ge-healthcare-inc/
• E.D. Tex. dockets (Xerox, Kyocera, Blu, TCL): Docket Alarm, RPX Insight, Dockets.Justia, PatSnap
• Joint termination/related-matters recital in IPR2019-00993 identifying IPR2019-01020 and the GE S.D. Cal. dismissal: https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1525541](/patent/1525541)/

If you want, I can next attempt to pull the actual PACER dockets for 2:21-cv-00345 and 2:22-cv-00360 to nail down their current statuses and confirm the '573 assertion counts.

I'll start by grounding on the one PTAB reference that appears in the patent record itself, then search for any others.

Key finding: an IPR exists in the record. Let me pin down the panel's reasoning and check for any other proceedings.

Let me try to retrieve the actual institution-denial reasoning and confirm no other proceedings exist.

Proceedings overview

One AIA trial proceeding exists on the record for US 7,092,573 — IPR2019-01020, filed by General Electric Co. and GE Healthcare, in which the Board denied institution on the merits; no claims were canceled, no claims were sustained by a Final Written Decision, no settlement occurred, and nothing was appealable. That is the entire PTAB footprint: 1 proceeding, 0 claims invalidated, 0 claims confirmed in an FWD, 1 institution denial.

Discrepancy flagged: the structured "PTAB proceedings on file" block states the USPTO ODP API returns no AIA trial proceedings for this patent. That is wrong (or at least stale). The Google Patents record for this patent itself carries a "Family has litigation" entry reading "PTAB case IPR2019-01020 filed (Not Instituted - Merits)", and third-party PTAB dockets confirm the case. I am treating IPR2019-01020 as on file. The ODP gap is likely an indexing artifact, but a defendant should confirm directly on PTAB E2E before relying on either source.

One attribution warning that matters: the Google Patents litigation block displays "Petitioner: 'Unified Patents PTAB Data' by Unified Patents is licensed under a Creative Commons Attribution 4.0 International License." That is a data-source license attribution for the PTAB dataset, not the identity of the petitioner. Unified Patents did not file this IPR. The petitioner was General Electric. Do not let a defensive aggregator assumption distort your read of the estoppel and privity analysis.

Bottom-line defensive posture: this is not a "hardened patent" story and not a "claims are canceled" story. It is a "never tested, but expired" story. All 33 claims of the '573 patent stand unadjudicated — the only merits look the Board ever gave them was a preliminary reasonable-likelihood screen that GE failed. The decisive defensive fact is not PTAB at all: per the patent record, the '573 patent carried an adjusted expiration of 2023-07-13 and is now "Expired - Lifetime." The assertion window is closed. Any defendant today is defending a past-damages-only case, with damages capped by the six-year lookback in 35 U.S.C. § 286.

IPR2019-01020 — General Electric Co. et al. v. Monument Peak Ventures, LLC

• Type: Inter Partes Review (35 U.S.C. §§ 311–319)

• Filed: 2019-04-25 (Petition, Paper 1; Notice of Filing Date Accorded to Petition mailed 2019-05-09)

• Status: Institution Denied (verbatim from the IP Verse PTAB case record). Google Patents renders the same outcome as "Not Instituted - Merits" — i.e., a merits-based denial, not a discretionary § 314(a)/§ 325(d) denial. Institution decision date: 2019-09-23.

• Judge panel: Not verified. The Paper 7 denial as surfaced in the sources available to me lists only the parties' counsel (Reed Smith LLP for Petitioner, Quinn Emanuel for Patent Owner). I could not confirm the APJ panel names and will not guess them. Check the caption page of Paper 7 on PTAB E2E.

• Petition grounds: Challenged claims 1, 2, 4–6, and 8 — that is, independent claim 1, dependent claims 2, 4, 5, 6, and independent claim 8 (the location-and-size variant). Statutory basis was § 103 obviousness, with the Petitioner's primary reference being US 7,092,569 to Kinjo ("Kinjo-569"). Petition exhibits show a Kinjo-centric combination/alternative-ground attack: Ex. 1004 US 7,092,569 (Kinjo), Ex. 1005 US 5,978,100 (Kinjo), Ex. 1008 US 5,754,348 (Sooho), Ex. 1009 US 4,800,379 (Yoemans), Ex. 1010 US 5,363,209 (Eschbach), plus non-patent literature Ex. 1006 (Luo et al., On the Application of Bayes Networks to Semantic Understanding of Consumer Photographs, ICIP 2000) and Ex. 1007 (Etz, Ground Truth for Training and Evaluation of Automatic Main Subject Detection, 2000). Expert: Alan C. Bovik (Ex. 1002 declaration, Ex. 1003 CV). Note the prior-art overlap with the patent's own file history: the Petitioner also submitted the 2006-02-09 Response and 2006-04-12 Notice of Allowance as Exs. 1014–1015, i.e., it was building a § 325(d)-style record about what the Examiner had already seen. I could not verify a ground-by-ground table or a ground count, so I am not stating one.

• Institution decision: DENIED — 2019-09-23 (Paper 7, "Decision Denying Institution of Inter Partes Review"). The Board ruled on the merits, not on discretion. The disposition language, quoted verbatim from the decision text:

  "For the foregoing reasons, we determine that Petitioner has not demonstrated a reasonable likelihood of prevailing on its challenges to any of claims 1, 2, 4-6, and 8."

  "ORDERED that the petition for inter partes review is denied and no trial is instituted."

  The Board's analysis walked claim 1's "belief map of values indicating the degree of belief that pixels in the digital image belong to target subject matter, said values defining a plurality of belief regions" element by element against Kinjo-569, and was evidently unpersuaded that Kinjo-569's "weight coefficients with a positive sign" / "negative sign" face-candidate weighting taught the claimed belief map. The decision also rejected Petitioner's position that the ordinarily skilled artisan would have understood image-enhancement decisions to be driven by subject-matter belief (the snippet I could retrieve cuts off mid-sentence, so I am not characterizing the full reasoning beyond this).

• Final Written Decision: None issued. No FWD exists, so no claim of US 7,092,573 has ever been canceled or confirmed at PTAB, and no claim-level patentability holding exists to cite.

• Settlement / termination: None. Patent Owner Monument Peak Ventures, LLC affirmatively waived its Preliminary Response (Paper 6, filed 2019-06-24), expressly reserving the right to file a Patent Owner Response and a motion to amend if trial were instituted. Trial never began, so no termination-on-settlement occurred and no settlement agreement exists (confidential or otherwise).

• Appeal: None, and none available. A denial of institution is a non-appealable, non-final Board action under 35 U.S.C. § 314(d) (see Cuozzo Speed Techs., LLC v. Lee, 136 S. Ct. 2131 (2016)). There is no Federal Circuit docket, and I found no request for rehearing. Consequently there is no PTAB estoppel to appeal and no CAFC decision interpreting any claim of this patent.

• Defensive value: The denial gives a defendant almost no estoppel benefit and almost no validity benefit. Because no trial was instituted, § 315(e)(2) estoppel never attached to GE or its privies — the prior art GE collected (Kinjo-569/Kinjo-100/Sooho/Yoemans/Eschbach) remains fully available to any defendant in district court, and GE itself is not barred from re-running it. Conversely, the petitioner's failure means the patent owner can point to a clean institution-stage loss as evidence that the claims are not facially weak. The real leverage for a defendant is elsewhere: the patent's expiration on 2023-07-13.

Related (not on this patent, but relevant to pattern): the same Petitioner filed a coordinated April-2019 cluster against Monument Peak Ventures' portfolio — IPR2019-00993 (US 6,509,910), IPR2019-01020 (this patent, US 7,092,573), IPR2019-01029 (US 7,212,668), and later IPR2019-01473 (US 6,654,507). In the '507 institution decision the Board expressly noted that "U.S. Patent No. 7,092,573 (which is the subject of IPR2019-01020) includes claim terms that may be relevant to claims of the '507 patent" and cited the denial: "General Electric Co. v. [Monument Peak Ventures], IPR2019-01020, Paper 7 (PTAB Sept. 23, 2019) (Decision Denying Institution)." These are different patents and are not proceedings on US 7,092,573 — do not count them.

Strategic summary

Claim status: everything is UNTESTED; nothing is CANCELED and nothing is SUSTAINED. Of the 33 claims, claims 1, 2, 4–6, and 8 were challenged in IPR2019-01020 and survived only a preliminary reasonable-likelihood screen — that is not a patentability adjudication. Claims 3, 7, 9–33 were never challenged at all. There is no surviving-claims list to report, because there is no FWD; do not let an opponent tell you the claims were "confirmed by the PTAB." They were not confirmed. They were also not canceled. The patent expired on 2023-07-13 with all claims intact and unadjudicated.

Estoppel landscape: there is essentially none, and that cuts both ways. Section 315(e)(2) estoppel is triggered by a petitioner that "in an inter partes review . . . that results in a final written decision" raised or reasonably could have raised a ground. IPR2019-01020 produced no final written decision, so no estoppel attached to GE, GE Healthcare, or their privies. Practical consequences for a defendant today: (1) the full Kinjo-569 + Kinjo-100 + Sooho + Yoemans + Eschbach + Luo/Etz art package is unencumbered and can be asserted in district court by anyone, including GE; (2) the Board's § 325(d)/Advanced Bionics-adjacent record GE built from the 2006 prosecution file history is equally available; (3) equally, the patent owner faces no adverse findings on any claim. The prior-art field is wide open — the constraint on a defendant is not estoppel, it is the date. Any new IPR would now be pointless as a practical matter, and anyone served with a complaint more than a year before filing is § 315(b)-barred in any event.

Pattern signals. GE is a large operating company, not a defensive aggregator, and it filed a portfolio-wide cluster against Monument Peak Ventures in a single three-week window in April 2019 driven by the S.D. Cal. litigation — one petition per asserted patent, only one petition against the '573 patent, and no follow-on petition after the denial. That is the signature of a litigation-driven, single-shot challenge, not a sustained invalidity campaign. Monument Peak Ventures is an Intellectual Ventures spin-out NPE (Eastman Kodak → Intellectual Ventures Fund 83 LLC, 2013-04-12 → Monument Peak Ventures, LLC, 2017-03-09) and litigated the patent owner side of IPR2019-01020 through Quinn Emanuel. Notably, the parallel S.D. Cal. case was stayed on 2019-05-13 pending institution decisions and then jointly dismissed on 2019-10-28/29 after the Board denied institution — the classic litigation-funded-IPR-then-dismiss pattern. The later assertion wave (E.D. Tex.: 2:22-cv-00360, 2:23-cv-00127, 2:23-cv-00169, and 5:24-cv-00011 against TCL, filed 2024-01-22) produced no PTAB filings at all — consistent with a patent that had already expired before those defendants were sued, leaving only backward-looking damages to argue about.

Recommended next steps

1. Anchor on expiration, not on PTAB. The '573 patent is recorded as Expired - Lifetime, adjusted expiration 2023-07-13 (https://patents.google.com/patent/US7092573/en). An expired patent cannot support injunctive relief. Any live exposure is money damages, and § 286 caps recovery at six years before the complaint — for a 2024 complaint that is roughly 2018-01 onward, clipped at the 2023-07-13 expiry. Build your damages model on that closed window first; it is worth more than any invalidity theory.
2. Do not tell a court or an adversary that the PTAB cleared or killed any claim. The accurate statement is: one IPR was filed, institution was denied on the merits on 2019-09-23, and no Final Written Decision ever issued. Quote the disposition if you need it: "For the foregoing reasons, we determine that Petitioner has not demonstrated a reasonable likelihood of prevailing on its challenges to any of claims 1, 2, 4-6, and 8" … "ORDERED that the petition for inter partes review is denied and no trial is instituted." Retrieve and cite IPR2019-01020, Paper 7 (PTAB Sept. 23, 2019) — via PTAB E2E (https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1525683](/patent/1525683)/download-documents) or the Docket Alarm case file (https://www.docketalarm.com/cases/PTAB/IPR2019-01020/Inter_Partes_Review_of_U.S._Pat._7092573/). The Unified Patents portal page (https://portal.unifiedpatents.com/ptab/case/IPR2019-01020) and IP Verse (https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2019-01020) both corroborate the denial and the 2019-09-23 date.
3. Verify the panel and the full Paper 7 text directly on PTAB E2E before filing anything that characterizes the Board's claim-1 reasoning. My reconstruction of the Board's rationale is partial — the retrieved decision text cuts off mid-analysis — and the APJ names are unconfirmed. If you cite the reasoning, cite the paper.
4. If a defendant is inside a one-year § 315(b) window and wants a reexam-grade kill, note the limits. A new IPR remains theoretically available to a non-time-barred petitioner even against an expired patent (claims construed under Phillips), but its only practical value is wiping out past damages — and the § 315(e)(2) downside of using it is now near zero, since you would be the only estopped party. More likely, ex parte reexamination or a district-court § 282 invalidity case on the untouched Kinjo/Sooho/Yoemans/Eschbach art is the cheaper path, especially given that no PTAB estoppel constrains anyone.
5. Confirm there is no additional PTAB activity I could not see. The ODP block says zero proceedings; Google Patents and third-party dockets say one. Given that contradiction, run a direct PTAB E2E search on "7,092,573" (and on application 10/016,601) to confirm the count of one before you represent to anyone that IPR2019-01020 is the patent's complete PTAB history. The E.D. Tex. assertion wave produced no PTAB filings, but I could not exhaustively verify that for every case, and the TCL action (5:24-cv-00011) is recent.

What I could not confirm, stated plainly: the APJ panel for IPR2019-01020; a ground-by-ground table (ground numbers, exact § 103 combinations per claim); whether a request for rehearing of the denial was filed; and the full text of the Board's claim-1 belief-map reasoning. I found no other IPR, PGR, or CBM naming US 7,092,573, no FWD, no settlement, and no Federal Circuit appeal.

I'll research the assignment record before building the output. Let me run several targeted searches.

Let me dig for the specific reel/frame records and litigation details.

Let me try to pin down reel/frame numbers and the correspondent of record.

I was unable to query the USPTO Assignment Center interface directly (it is a dynamic application that does not render to the crawler), and my access to further web lookups is now exhausted. I reconstructed the chain from the Google Patents legal-events record, the underlying assignment document text surfaced via a PTAB exhibit, and the litigation/SEC record. Reel/frame numbers and the recording correspondents are NOT available to me in this session — I have left those fields explicitly marked rather than guess.

Inventors

All four appear as assignors on the 2001-12-10 assignment to Eastman Kodak Company, which is the only direct evidence of their employer at filing. The specification is Kodak image-science work (subject-matter belief maps for flesh, sky, grass, snow, water; Huber‑MRF JPEG de-blocking), consistent with the Kodak Research Labs image-processing group.

Pattern note / caution: I cannot substantiate the brief's "all inventors depart within 12 months of filing" hypothesis for this patent, and I will not assert it. One caveat worth logging: the inventor name Amit Singhal collides with the well-known Google search engineer, while non-patent citation "Singhal et al., A Multilevel Bayesian Network Approach to Image Sensor Fusion, Int'l Conf. on Information Fusion, July 2000" is an image-fusion paper consistent with a Kodak imaging researcher. I flag this as an identity question to verify, not as a finding.

Original assignee

Eastman Kodak Company — the sole assignee on the face of US 7,092,573 (filed and granted 2001-12-10 / 2006-08-15).

• Line of business: consumer and professional imaging — film, cameras, digital cameras, photofinishing, and (from the 1990s) digital image processing/science. Kodak was a pioneer here: the first self-contained digital camera (1975) and the first megapixel sensor (1986), per MPV's own complaint (Monument Peak Ventures, LLC v. GE Healthcare, Inc., S.D. Cal. 3:18-cv-01158, ¶¶11–13).
• Did they ship a product embodying the claims? Kodak plausibly practiced content-adaptive enhancement in its digital-imaging products and licensing programs — its patent licensing revenue reached ~$838M in 2010 (per the same complaint, ¶13). However, I cannot tie the specific claims of the '573 patent to a named Kodak product from the sources retrieved; treat "product embodiment" as unverified.
• Current status: Bankrupt and reorganized. Kodak filed Chapter 11 in January 2012 (Bankr. S.D.N.Y.); it sold ~1,100 digital-imaging patents to Intellectual Ventures Fund 83 LLC for $527M (announced Dec 2012; closing ~Feb 1, 2013); emerged from Chapter 11 in 2013. The '573 patent is now expired (adjusted expiration 2023-07-13), expired-lifetime status.

Assignment timeline

Data-integrity note: the USPTO's recorded reel/frame numbers and the correspondent of record on each cover sheet were not retrievable in this session. The events below are real recorded/legal events (they appear in the Google Patents legal-events feed sourced from USPTO assignment data, and in one case I retrieved the underlying assignment instrument), but I have not fabricated reel/frame or correspondent values.

• 2001-12-10 (executed) / recorded 2001-12-10 — Reel not retrieved

  • Conveyance: Assignment of inventors' interest
  • Assignor: Amit Singhal; Andrew C. Gallagher; Robert T. Gray; Jiebo Luo
  • Assignee: Eastman Kodak Company
  • Correspondent: not retrieved
  • Context: Original employment assignment; all four inventors conveyed to Kodak at filing.

• 2012-02-21 (recorded) — Reel not retrieved

  • Conveyance: Security Agreement (security interest)
  • Assignor: Eastman Kodak Company (with Pakon, Inc.)
  • Assignee: CITICORP NORTH AMERICA, INC., AS AGENT
  • Correspondent: not retrieved
  • Context: Securitization — Kodak's post-Chapter-11 DIP/term-loan collateral package; the '573 patent pledged as collateral.

• 2013-02-01 (recorded) — Reel not retrieved

  • Conveyance: Patent Release
  • Assignor: Citicorp North America, Inc. and Wilmington Trust, National Association (as agents)
  • Assignee: Kodak entities (release)
  • Correspondent: not retrieved
  • Context: Internal/release — lien released as Kodak repaid term debt out of the patent-sale proceeds (~$418.7M applied to term loans per Kodak's Q1 FY2013 Form 8-K).

• 2013-04-12 (recorded) — Reel not retrieved

  • Conveyance: Assignment (patent sale)
  • Assignor: Eastman Kodak Company
  • Assignee: INTELLECTUAL VENTURES FUND 83 LLC
  • Correspondent: not retrieved
  • Context: Bankruptcy fire-sale. Sale of the ~1,100-patent digital-imaging portfolio for $527M under the Sale Agreement approved by the Bankruptcy Court; funded by an IV/RPX-organized licensee consortium (Adobe, Amazon, Apple, Facebook, FUJIFILM, Google, HTC, Huawei, Microsoft, RIM, Samsung, Shutterfly); Kodak retained a grant-back license.

• 2017-02-15 (executed) / recorded 2017-03-09 — Reel not retrieved

  • Conveyance: Assignment of Patent Rights (portfolio sale)
  • Assignor: Intellectual Ventures Fund 83 LLC, a Delaware LLC, 7251 W Lake Mead Blvd, Ste 300, Las Vegas, NV 89128
  • Assignee: MONUMENT PEAK VENTURES, LLC, a Texas LLC, 1400 Preston Road, Suite 400, Plano, TX 75093
  • Correspondent: not retrieved
  • Context: Transfer to asserter. The instrument conveys the patents plus "all causes of action, enforcement rights and all other rights to seek and obtain any other remedies … for past, current and future infringement" — i.e., expressly assigned for enforcement. (I retrieved the text of this IV‑Fund‑83→MPV assignment instrument; its Appendix lists a Kodak-origin portfolio.)

• 2023-08-15 (recorded) — Reel not retrieved

  • Conveyance: Release by Secured Party
  • Assignor: Intellectual Ventures Fund 83 LLC
  • Assignee: Monument Peak Ventures, LLC
  • Correspondent: not retrieved
  • Context: Internal/release — clearing a security interest of record; occurs after the patent's adjusted expiration (2023-07-13) and contemporaneous with the 2023 Texas suits.

Litigation overlay (not assignments, but drives the timeline):

• 2018-06-04 — Monument Peak Ventures, LLC v. GE Healthcare, Inc., No. 3:18-cv-01158 (S.D. Cal.), asserting the '573 patent (claims 1, 2, 4–6, 8) plus the '910 and '668 patents. MPV's corporate disclosure identifies Dominion Harbor Group, LLC as its parent.
• 2019-04-25 — IPR2019-01020, General Electric Co. v. Monument Peak Ventures, LLC (U.S. 7,092,573); institution denied 2019-09-23.
• 2022–2024 — E.D. Tex. suits naming MPV: 2:22-cv-00360; 2:23-cv-00127; 2:23-cv-00169; and Monument Peak Ventures, LLC v. TCL Electronics Holdings Ltd., 5:24-cv-00011 (filed 2024-01-22).

Timeline diagram

timeline
    title Ownership of US 7092573
    2001 : Patent filed by Eastman Kodak
         : Inventors assign to Kodak
    2012 : Kodak pledges patent to Citicorp as agent
    2013 : Lien released
         : Kodak sells portfolio to Intellectual Ventures Fund 83
    2017 : IV Fund 83 transfers patent to Monument Peak Ventures
    2018 : Monument Peak Ventures sues GE Healthcare
    2019 : GE files IPR against the 573 patent
         : IPR institution denied
    2023 : Patent reaches adjusted expiration
         : Secured party release recorded
    2024 : Monument Peak Ventures sues TCL in Texas

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT. The operating assignee (Eastman Kodak) transferred to a licensing special-purpose vehicle, Intellectual Ventures Fund 83 LLC (a Delaware LLC holding sold Kodak imaging patents), recorded 2013-04-12, and that fund in turn conveyed to Monument Peak Ventures, LLC, a Texas LLC at 1400 Preston Road, Suite 400, Plano, TX 75093 (executed 2017-02-15, recorded 2017-03-09). MPV's own corporate disclosure names Dominion Harbor Group, LLC as parent, and GE's brief called MPV "a patent assertion entity." The assignment instrument expressly conveys past/future infringement causes of action. The suffix "Ventures," no products in commerce, and a single-purpose holding LLC are all satisfied.

2. Known asserter in the chain — PRESENT (strong). Two links: Intellectual Ventures (a listed high-frequency monetization entity) as the 2013–2017 owner, and Monument Peak Ventures / Dominion Harbor Enterprises as current owner. Dominion Harbor is a publicly identified patent-monetization/assertion firm; MPV's board includes the asserter pattern (serial plaintiffs across E.D. Tex. 2022–2024). This is corroborated by an opposing party's on-record characterization, not naming alone.

3. Repeat correspondent across the chain — UNCLEAR / NOT ESTABLISHED. The correspondent of record on each recorded cover sheet is the single most diagnostic field for this signal, and I could not retrieve it. What is visible is recurring litigation counsel for MPV (e.g., Quinn Emanuel — Amar Thakur; Bruce Zisser), which is assertion-side representation, not assignment-recording correspondence. Do not count this signal on the evidence available; re-check the Assignment Center cover sheets for the 2013 and 2017 recordings.

4. Cascading transfers (<24 months through chained LLCs) — NOT PRESENT. The chain is Kodak → IV Fund 83 (2013) → MPV (2017): the IV→MPV step is ~4 years after the Kodak sale, and each link is a distinct transaction, not a rapid relay. No evidence of same-day or back-to-back LLC hops.

5. Pre-litigation transfer (assignment within 6 months before first suit) — NOT PRESENT / UNCLEAR. The IV→MPV transfer executed 2017-02-15 (recorded 2017-03-09); the earliest suit naming this patent that I can confirm is the GE action filed 2018-06-04 — roughly 15 months later. If an earlier MPV complaint (2017) exists, this would flip to present; I could not confirm one.

6. Bankruptcy fire-sale — PRESENT (strong). Eastman Kodak's Chapter 11 (filed Jan 2012) produced the $527M sale of the digital-imaging portfolio to IV Fund 83, recorded 2013-04-12; Kodak applied ~$418.7M of proceeds to term loans. This is a textbook bankruptcy-driven patent sale (cf. Nortel, Polaroid).

7. Privateering — UNCLEAR. The Kodak sale was structured with a licensee consortium organized by IV and RPX — arguably defensive aggregation of the portfolio against outside assertion. But the downstream transfer of the '573 patent to MPV (Dominion Harbor), an asserter with no manufacturing business, means the portfolio was ultimately monetized offensively against third parties (GE, TCL) rather than used purely defensively. I would not mark this "present" as classic privateering (Kodak did not commission suits against competitors), but the defensive intent of the original consortium did not survive for this patent.

8. Defensive aggregator endpoint — NOT PRESENT. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at an asserting entity. (Inverse of the neutralized-patent signal.)

Verdict

NPE — high confidence (3 strong signals: #1 shell-entity transfer, #2 known-asserter/IV + Dominion Harbor, #6 bankruptcy fire-sale).

The chain is Eastman Kodak → Intellectual Ventures Fund 83 LLC (reel not retrieved, recorded 2013-04-12, bankruptcy fire-sale) → Monument Peak Ventures, LLC (executed 2017-02-15, recorded 2017-03-09, Texas LLC, parent Dominion Harbor Group). The 2013 step is a Kodak Chapter 11 asset sale, and the 2017 step is a patent-monetization special purpose vehicle that expressly took the past-infringement causes of action and then sued GE Healthcare (2018, 3:18-cv-01158), weathered IPR2019-01020 (institution denied), and filed multiple E.D. Tex. suits through 2024 — the signature of an assertion-driven NPE chain, not an operating company.

Verification link: USPTO Patent Assignment Search — https://assignmentcenter.uspto.gov/ (search by patent number 7092573), and the legacy interface at https://assignment.uspto.gov/patent/index.html. Please re-pull the reel/frame and correspondent fields there, as those are the only fields I could not independently obtain and the only ones that could upgrade signal #3 (repeat correspondent).

I'll research this patent and its cited prior art.

I have the full patent and its citation list. Let me verify key prior-art references and the PTAB proceeding.

Prior-Art Analysis — US 7,092,573

0. Patent identified (verified by search)

Independent claims are 1, 7, 8, 9 and 13. The core concept across them is: (a) a subject-matter detector producing a belief map of degrees of belief, and (b) pixel-by-pixel variation of the enhancement according to the belief map and — in claims 1, 7, 8, 9 — the size and/or location of the belief regions.

1. Important framing before the claim-mapping

The question asks which claim(s) each cited reference "potentially anticipates under 35 U.S.C. § 102." Two honest caveats must be stated up front:

1. Strict § 102 anticipation requires a single reference disclosing every limitation. None of the 26 references cited on the face of the '573 patent discloses the full combination of the independent claims — i.e., a belief map of graded belief values whose region size (and/or location) modulates a pixel-by-pixel enhancement control signal. The references below are better characterized as element-level § 102 art and as the obviousness (§ 103) base. I flag where a reference comes closest.
2. The reference dates below are taken from the '573 patent's own front-page "Patent Citations (26)" list (as reproduced on Google Patents); I did not independently re-verify each document's text beyond the searches noted. This is an analytical summary, not a legal opinion or a validity conclusion.

2. Patent citations, with § 102 mapping

Tier A — Directly on-point (enhancement controlled by detected subject matter)

A1. US 5,682,443 — Gouch et al. (Crosfield Electronics), "Image color modification method and apparatus employing unsharp masking"

• Priority 1993-03-24 (GB 9306083/9308794); filed 1994-03-21; published/granted 1997-10-28.
• Description: modifies the unsharp-mask fringe signal pixel-by-pixel with a weighting function φ derived from the sharp/unsharp color values, so the degree of sharpening/smoothing is set by the color of the region (their stated example: sharpen hair, smooth skin). This is the reference the '573 specification singles out as the closest color-based approach.
• § 102 relevance: The strongest single-reference attack on the sharpening-specific dependent claims — e.g., claim 23 (enhancement is sharpening) read with the pixel-by-pixel limitation of claim 1/13. It does not anticipate the independent claims because it uses color only, with no belief map, no graded probability, and no size/location dependence (the '573 specification itself draws this distinction).

A2. EP 0 681 268 A1 — "Automatic image correction using pattern recognition techniques" (Eastman Kodak)

• Filed 1994-04-22; published 1995-11-08.
• Description: a pattern analyzer (neural network / eigenvector-template matching) detects the presence and location of "color-significant objects" (e.g., a human face); a transform generator applies rules such as "if sharpening an image, do not sharpen faces" and derives color/sharpness transforms applied to the image.
• § 102 relevance: Closest art on the "detect subject matter → generate control transform → apply enhancement" architecture of independent claims 9 and 13. It discloses detection + location + selective sharpening, but the detection output is an object classification, not a graded belief map, and there is no size-dependent control — so it reads on the preamble/steps of 9/13 but not the full combination.

A3. US 2003/0108245 A1 — Luo (Eastman Kodak), "Method and system for improving an image characteristic based on image content"

• Priority/filed 2001-12-07; published 2003-06-12.
• Description: sibling Kodak application (same inventor family) that improves an image characteristic (e.g., sharpening) based on detected image content.
• § 102 relevance: Only potentially § 102(e) art if treated as of its 2001-12-07 filing date (one day before the '573 priority). It is highly relevant to claims 9/13, but because it is essentially co-pending/co-invented sibling subject matter, expect it to be addressed as § 102(e)/§ 103 rather than clean anticipation.

Tier B — Subject-matter detection / belief-map generation (feed the "belief map" element)

B1. US 5,901,245 — Warnick, Senn & Mehrotra (Eastman Kodak), "Method and system for detection and characterization of open space in digital images"

• Priority 1997-01-23; filed 1997-01-23; granted 1999-05-04.
• Description: builds an activity map and identifies contiguous low-activity (smooth) regions as "open space." This is the open-space detection step 1022/1024 embedded in the '573 FIG. 10 and expressly incorporated by reference in the '573 specification.
• § 102 relevance: Element-level art for the belief-region / contiguous-region-forming limitation (background of claims 1/7/8/9). Does not address enhancement control, so not an anticipation of any independent claim.

B2. US 6,504,951 B1 — Luo et al. (Eastman Kodak), "Method for detecting sky in images"

• Priority 1999-11-29; granted 2003-01-07.
• Description: creates a belief map for blue sky (the Ser. No. 09/450,190 work cited in the '573 specification).
• § 102 relevance: Element-level art for the "belief map" step and specifically for the clear-blue-sky target of claim 19 and the multi-value belief limitation of claim 33.

B3. US 5,710,833 — Moghaddam et al. (MIT), "Detection, recognition and coding of complex objects using probabilistic eigenspace analysis"

• Priority 1995-04-20; granted 1998-01-20.
• Description: probabilistic (eigenspace) detection of faces/objects — cited in the '573 specification for face detection.
• § 102 relevance: Element-level art for the belief/probability-based detection and for the human-face/flesh target of claim 18.

B4. US 4,203,671 — Takahashi et al. (Fuji Photo Film), "Method of detecting flesh color in color originals"

• Priority 1976-06-22; granted 1980-05-20. Cited in the '573 specification as a skin-detection reference whose likelihood-only approach the patent criticizes.
• § 102 relevance: Very old element-level art for human-flesh detection (claim 18). Does not touch enhancement control.

B5. US 5,781,276 — Zahn et al. (Agfa-Gevaert), "Printing of color film"

• Priority 1992-07-27; granted 1998-07-14. Skin-tone detection via P(color|Skin).
• § 102 relevance: Same role as B4 for the skin/flesh target (claim 18).

B6. US 2002/0076100 A1 — (Eastman Kodak), "Image processing method for detecting human figures in a digital image"

• Priority 2000-12-14; published 2002-06-20.
• § 102 relevance: Element-level art for human-figure detection (claim 18) and for the "apply a subject-matter detector" step.

B7. EP 1 017 019 A2 — (Eastman Kodak), "Method for automatic determination of main subjects in photographic images"

• Priority 1998-12-31; published 2000-07-05.
• § 102 relevance: Element-level art for main-subject / region determination feeding a belief map (claims 1, 7, 8, 9).

B8. US 5,889,578 — (Eastman Kodak), "Method and apparatus for using film scanning information to determine the type and category of an image"

• Priority 1993-10-26; granted 1999-03-30.
• § 102 relevance: Background art for classifying image type/category prior to processing — tangential to claims 1/13.

Tier C — Enhancement operations recited in the dependent claims

C1. US 6,636,645 B1 — (Eastman Kodak), "Image processing method for reducing noise and blocking artifact in a digital image"

• Priority 2000-06-29; granted 2003-10-21.
• § 102 relevance: Element-level art for the noise reduction and JPEG de-blocking dependent claims (24, 28).

C2. US 6,891,977 B2 — (Eastman Kodak), "Method for sharpening a digital image without amplifying noise"

• Priority 2002-02-27; granted 2005-05-10.
• § 102 relevance: Sharpening art relevant to claim 23. Date anomaly flagged: its 2002-02-27 filing date post-dates the '573 2001-12-10 priority date, so it cannot be § 102(e) prior art as to the '573 claims unless the inventors' actual invention date is later; treat its citation as § 103/background or as a listed-but-non-qualifying reference.

C3. US 6,845,181 B2 — (Eastman Kodak), "Method for processing a digital image to adjust brightness"

• Priority 2001-07-12; granted 2005-01-18.
• § 102 relevance: Element-level art for the scene-balance adjustment dependent claim (26).

C4. US 5,523,849 — (Eastman Kodak), "Optimizing edge enhancement for electrographic color prints"

• Priority 1993-06-17; granted 1996-06-04.
• § 102 relevance: Background art for edge enhancement (claim 23).

C5. US 5,848,181 — (Sony), "Image processing method, image processing apparatus, noise removing method, and noise removing apparatus"

• Priority 1995-07-26; granted 1998-12-08.
• § 102 relevance: Element-level art for noise removal (claim 24).

C6. US 4,682,230 — (RCA), "Adaptive median filter system"

• Priority 1986-03-21; granted 1987-07-21.
• § 102 relevance: Very old art for adaptive/noise filtering (claim 24).

C7. US 6,141,017 — Cubillo et al. (Iterated Systems), "Method and apparatus for scaling an array of digital data using fractal transform"

• Priority 1998-01-23; granted 2000-10-31. Cited in the '573 specification as the fractal-based interpolation option.
• § 102 relevance: Element-level art for the image magnification / fractal interpolation dependent claims (29, 30).

C8. US 6,731,821 B1 — (Hewlett-Packard), "Method for enhancing compressibility and visual quality of scanned document images"

• Priority 2000-09-29; granted 2004-05-04.
• § 102 relevance: Background art for compression-artifact / quality enhancement; tangential.

C9. US 6,665,446 B1 — (Canon), "Image processing apparatus and method"

• Priority 1998-12-25; granted 2003-12-16.
• § 102 relevance: General image-enhancement background; tangential.

C10. US 6,453,069 B1 — (Canon), "Method of extracting image from input image using reference image"

• Priority 1996-11-20; granted 2002-09-17.
• § 102 relevance: Background art for region/object extraction (belief-region formation).

C11. US 6,707,940 B1 — (Intel), "Method and apparatus for image segmentation"

• Priority 2000-03-31; granted 2004-03-16.
• § 102 relevance: Background art for image segmentation — supports the "belief region" element (claims 1, 7, 8).

C12. US 6,738,494 B1 — (Eastman Kodak), "Method for varying an image processing path based on image emphasis and appeal"

• Priority 2000-06-23; granted 2004-05-18.
• § 102 relevance: Element-level art for the "select one or more image enhancement operations" limitations (claims 6, 12, 16) and for content-driven processing-path selection.

Tier D — Cited but of limited/incidental relevance

D1. US 2001/0031804 A1 — (General Electric), "Color stable compositions containing arylate-comprising polymers" — priority 1999-09-10; pub. 2001-10-18. Polymer-chemistry subject matter; appears unrelated to image processing and is likely an incidental/IDS citation. No § 102 relevance to any '573 claim.

D2. US 2003/0093279 A1 — Malah, "System for bandwidth extension of narrow-band speech" — priority 2001-10-04; pub. 2003-05-15. Speech signal processing; likely cited for a signal-processing technique (e.g., spectral/filter estimation), not for image semantics. No meaningful § 102 mapping.

D3. US 2002/0093670 A1 — (Eastman Kodak), "Doubleprint photofinishing service with the second print having subject content-based modifications" — priority 2000-12-07; pub. 2002-07-18. Content-based modification background (claims 13/17 generally).

3. Non-patent citations (11) — also material

These are the references the examiner listed; several are more directly relevant than many of the patents:

Note: the two Luo/Dupin applications are co-owned, co-pending Kodak applications — they are the source of the belief-map detectors used by the '573 invention and are best treated as § 102(e)-type or background art rather than anticipating references.

4. Additional art surfaced by the PTAB record (context beyond the front page)

In IPR2019-01020, the petitioner (General Electric / a Monument Peak Ventures challenger) relied on additional references, including US 7,092,569 to Kinjo and the following NPL, which a validity analyst should consider alongside the front-page citations:

• Luo, J. et al., "On the Application of Bayes Networks to Semantic Understanding of Consumer Photographs," ICIP, Sep 13, 2000.
• Etz, S.P., "Ground Truth For Training and Evaluation of Automatic Main Subject Detection," Human Vision and Electronic Imaging, Jun 2, 2000.
• Ekin, A. et al., "Extraction of Semantic Description of Events Using Bayesian Networks," ICASSP, May 11, 2001.

The petition was not instituted, which is consistent with the view that the independent claims' belief-map-plus-size/location control combination was not shown to be anticipated or obvious on that record.

5. Bottom line — most relevant prior art and claim exposure

Most likely to be asserted as anticipating under § 102: the sharpening-specific dependent claims (notably claim 23, and the pixel-varying aspect of claims 1/13) against US 5,682,443 (Gouch); and the detect-and-selectively-enhance flow of claims 9/13 against EP 0 681 268 A1.

Least likely to be asserted at all: US 2001/0031804 A1 (arylate polymers) and US 2003/0093279 A1 (speech bandwidth extension) — incidentally cited and not probative of any '573 claim.

No single cited reference anticipates independent claims 1, 7, 8, 9 or 13, because every one of them lacks the feature the '573 claims make central — using the size (and, in claims 8/9, the location) of belief regions to modulate enhancement on a pixel-by-pixel basis. If validity is challenged, the realistic theory is a § 103 combination (e.g., Gouch's pixel-varying sharpening + EP 0 681 268's object detection + a belief-map generator such as Luo's sky/face detectors), not clean anticipation.

I'll ground the reference details with searches before building the combinations.

Good grounding on the primary references. Now let me check the region-size/main-subject references and the PTAB record.

Obviousness Analysis — U.S. Patent 7,092,573 B2

"Method and system for selectively applying enhancement to an image"
Luo, Gallagher, Singhal & Gray; filed/priority 2001‑12‑10; granted 2006‑08‑15; original assignee Eastman Kodak Co.; current assignee Monument Peak Ventures LLC. Source: https://patents.google.com/patent/[US7092573](/patent/US7092573)/en (full text PDF: https://patentimages.storage.googleapis.com/pdfs/US7092573.pdf)

Scope note: this is a technical analysis of the record on the patent page, not a legal opinion. Any actual invalidity holding requires claim construction and evidentiary proof.

1. Governing framework

The application was filed 2001‑12‑10, so pre‑AIA 35 U.S.C. § 103 governs (Graham v. John Deere, 383 U.S. 1 (1966)). The KSR rationales (550 U.S. 398 (2007); MPEP 2144) apply to the motivation analysis: combining known elements according to known methods to yield predictable results; simple substitution of one known element for another; use of a known technique to improve a similar device in the same way; and "obvious to try" where the prior art presents a finite number of identified, predictable solutions.

Level of ordinary skill (proposed): a B.S. in EE/CS/imaging science (or equivalent) plus ~2–3 years of digital image processing experience, including familiarity with unsharp masking, skin/sky color classification, and probabilistic ("belief map") image segmentation. The references below are drawn from exactly that body of work and are largely commonly assigned or cited in the same art.

Threshold qualification of the art. Under pre‑AIA § 103(c), references that qualify only under § 102(e) and were commonly owned at the time of invention are disqualified as § 103 art. That matters here because most of the most on‑point references are Kodak's own (the 573 patent is a Kodak filing). The combination proposed in § 4 below is therefore deliberately anchored on § 102(b) art (printed publications and patents published more than one year before 2001‑12‑10), so it does not depend on the commonly‑owned 102(e) material:

⚠ Date caveats to verify: (i) the Cho 2001 article must be shown to predate 2001‑12‑10 to be § 102(b) art — otherwise it is § 102(a) art limited by the invention date. (ii) The page contains internal discrepancies in the identifier strings that should not be auto‑corrected: the Description cites Cho at "34 (2001) pp. 1067–1073" while the Non‑Patent Citations list says "pp. 1067‑1083"; the Description cites Lee at "Vol. 24, pp. 189–198, April 1983" while the NPL list says "pp. 255‑269"; and the Description cites "U.S. Ser. No. 09/904,366 filed Jul. 21, 2001 Dupin et al." while the NPL list says "filed Jul. 12, 2001." These must be reconciled against the physical documents before any petition.

EP prosecution admission. The European counterpart record (EP 1 317 475 B1) states that Singhal et al. (Fusion 2000) "teaches a subject matter detector for distinguishing between target and background subject matters and applying the subject matter detector to the image to produce a belief map indicating the degree of belief that pixels in the image belong to target subject matter." https://patentimages.storage.googleapis.com/12/d9/98/a88162e45e0395/EP1318475B1.pdf This is a direct, on‑the‑record identification of the belief‑map element.

2. Distilled claim requirements

Independent claims 1, 7, 8, 9, 13 (and product claim 32) all share:

1. A subject matter detector applied to the image →
2. a belief map of values (claim 2/33: three or more / more than two different values) →
3. the values define belief regions →
4. region size (claims 1, 7, 13) and/or region location (claims 8, 9) are determined →
5. an enhancement whose strength varies pixel by pixel in dependence on belief and region size (and/or location).

Claim 3 adds color and texture classification; claims 4–6, 10–12 add selection from plural detectors/operations and resolution reduction; claims 14–30 enumerate target subject matters and enhancement operations; claims 29–30 add interpolation (bilinear vs. fractal).

The only genuinely atypical limitation is item 4 coupled with item 5 — i.e., region‑level attributes (size, location) feeding a pixel‑level enhancement control signal. Everything else was squarely known by December 2001.

3. Prior‑art groupings

Group A — per‑pixel, content‑dependent enhancement control

• US 5,682,443 (Gouch) — unsharp masking where "the fringe amplitude (ie. degree of sharpening) is varied according to the colour of the region… e.g. sharpening hair and smoothing skin colours." Verified: the weighting function φ is set per pixel from the sharp/unsharp colour values; the specification states the degree of sharpening "may be set automatically according to some predetermined rule set," and the patent notes the pre‑existing approach required manual masking/retouch. https://patents.google.com/patent/[US5682443A](/patent/US5682443A)/en
• EP 0 681 268 A1 (Schwartz/Kodak) — pattern analyzer detects "the presence and location of color‑significant objects" (e.g., a human face), and a transform generator derives color and sharpness transforms, with the express example rule: "If sharpening an image, do not sharpen faces." https://patents.google.com/patent/EP0681268A1/en

Group B — probabilistic belief maps / subject‑matter detection

• Singhal et al., Fusion 2000 (belief map from Bayesian sensor fusion) — identified as such in the EP prosecution.
• Cho et al. (2001) / US 4,203,671 (Takahashi) / US 5,781,276 (Zahn) — graded skin‑tone probability; the 573 specification itself notes Takahashi/Zahn use likelihood P(color|Skin) and that a Bayes‑corrected conditional probability is preferable, i.e. the graded‑belief approach was known.
• US 5,710,833 (Moghaddam) — probabilistic eigenspace face detection.
• Saber et al. (1996) — color classification for sky, expressly criticized in the 573 background for false positives (clothing, man‑made surfaces, water) — this is the problem statement that motivates belief‑based classifiers using texture/shape as well as color.

Group C — region extraction, size and location characterization

• US 5,901,245 (Warnick) — identifies "contiguous pixels of the activity map having values below the predetermined threshold" as regions, and produces metadata including "extent," "openness," location, color and texture. Verified: https://patentimages.storage.googleapis.com/2a/30/46/977c45825b8da9/US5901245.pdf
• EP 1 017 019 A2 / US 6,282,317 (Luo, Etz, Singhal) — extracts "regions of arbitrary shape and size," computes structural saliency features including size (and size contrast, centrality/location PDF, borderness), and integrates them into "a belief that each region is the main subject," output as "a map in which the brightness of a region is proportional to the main subject belief of the region." https://patents.google.com/patent/EP1017019A2/en This is the single most important reference for the size/location elements.

Group D — the enhancement primitives named in the claims

• Sharpening: Jain (pp. 249–250); Luo US 6,891,977.
• JPEG de‑blocking: Luo et al., IEEE TIP Vol. 5 No. 9, Sept. 1996, pp. 1363–1368; US 6,636,645; HMRF/Huber potential (as described in the 573 spec itself).
• Noise reduction: Lee σ‑filter.
• Brightness/scene balance: US 6,845,181.
• Interpolation: US 6,141,017 (Cubillo, fractal scaling) and bilinear (Jain).
• Belief‑map‑driven sharpening modification: US 2002/0093670 A1 (Luo & Gray, Kodak) — main‑subject belief map used to alter the background "with reduced sharpness" (102(e) art with § 103(c) caveat).

4. The primary obviousness combination

Combination 1 — Claims 1, 2, 7, 33 (the core: size‑sensitive enhancement)

Gouch (US 5,682,443) + EP 0 681 268 A1 + Singhal/Fusion 2000 + EP 1 017 019 A2 (US 6,282,317), optionally + Cho/Saber.

Motivation to combine (rationale‑by‑rationale):

1. Same field, same problem, predictable result (MPEP 2144.03): all references address the same defect — global, one‑size‑fits‑all enhancement degrading faces/sky or destroying texture. Gouch supplies the per‑pixel control knob; the detector references supply the semantic input. Output is the predictable improvement of each.
2. Express "known technique" suggestion: Gouch itself says the weighting function "may be set automatically according to some predetermined rule set," and laments that region‑selective sharpening "cannot be included in the scanning process" without manual spatial masks. Replacing the manual mask with an automated detector (EP 0 681 268's rule "do not sharpen faces"; EP 1 017 019's belief map) is the very improvement Gouch invites.
3. Gouch's own admitted deficiency supplies the gap‑filling motivation: the 573 background asserts that Gouch's color‑only approach "likely unnecessarily conservatively sharpens image regions having a similar color to human flesh such as bricks or wood" and "does not allow for the use of other features such as texture or shape features." Saber et al. makes the identical criticism of color‑only sky detection (misclassifying clothing, man‑made surfaces, water). A POSITA reading Gouch with Saber/Cho/EP 1 017 019 would have been motivated, with a reasonable expectation of success, to substitute a multi‑feature probabilistic classifier for the raw color/keying weighting function. That is a simple substitution of one known element (color lookup) for another (belief value) to obtain a predictable improvement — a textbook KSR rationale.
4. Admitted prior art / stated problem in the 573 specification: the Background acknowledges that global enhancement over‑sharpens faces and under‑sharpens grass, and that region‑by‑region adjustment was previously an "expensive" manual process. Such statements of the then‑existing problem and its known solution direction weigh strongly in the § 103 analysis (cf. In re Nomiya; Riverwood Int'l v. R.A. Jones).
5. No change in principle of operation: Gouch's β(x,y)‑style per‑pixel multiplier and the 573 claim's pixel‑wise control signal are functionally identical hardware/software constructs; the modification amounts to reprogramming the control input.

Combination 2 — Claims 8 and 9 (location + size)

Combination 1 + EP 1 017 019 A2 / US 6,282,317 (centrality: "the main subject tends to be located near the center…", location PDF over the region's pixels) and/or US 5,901,245 ("the location of the region relative to the entire image"). The 573's own FIG. 8 belief‑map analyzer merely outputs an L(x,y) "distance from each pixel location to the center of the image or some other appropriate distance" — a distance‑to‑center map is precisely the centrality construct of EP 1 017 019, making claims 8/9 largely a matter of applying a known saliency feature to a known control signal.

Combination 3 — Claims 4–6, 10–12, 14–31 (detector/operation selection and enumerations)

These are enumerations of known primitives, each independently known and combinable with Combination 1 without any change in principle:

• Sharpening (claim 23): Gouch + Jain + US 6,891,977; US 2002/0093670 A1 goes further and applies a main‑subject belief map to alter background sharpness.
• Noise reduction (claim 24): Lee σ‑filter (the 573 spec itself describes the σ‑filter as the state of the art).
• Tone scale / scene balance (claims 25, 26): US 6,845,181 + EP 0 681 268 (transform generation from semantic content).
• Color re‑mapping (claim 27): Gouch's color‑space weighting + the well‑known "memory color" concept.
• JPEG de‑blocking (claim 28): Luo et al., IEEE TIP 1996 + US 6,636,645 + HMRF/Huber potential (described in the 573 spec as known).
• Interpolation (claims 29–30): US 6,141,017 (fractal) + bilinear (Jain); the selection between planar and non‑planar models is an ordinary design choice.
• Resolution reduction (claims 5, 31): conventional subsampling, as described in US 2003/0108245 A1 and acknowledged in the 573 spec ("Many methods… are known in the art… see Gonzalez et al.").
• Target subject matters (claims 15, 18–22): Gouch (flesh); Luo Ser. No. 09/450,190 and US 5,901,245 (sky/open space); EP 1 017 019 (people/sky/grass key subject matters); snow and water by analogous detectors, as the 573 spec concedes.
• Computer program product (claim 32): routine implementation (the 573 spec lists conventional storage media).

5. Where the challenger is strongest — and weakest

Strongest ground: independent claims 1, 7, 13 on the § 102(b) combination Gouch + EP 0 681 268 + EP 1 017 019 A2 (+ Singhal). Every element has an express textual hook, and two of the references contain self‑executing motivation language (Gouch's automatic rule‑set weighting; EP 0 681 268's "do not sharpen faces" rule).

Weakest link — the "size" limitation. Claim 1 requires that the enhancement vary "in accordance with both the degree of belief **and the size of the respective said belief region." Gouch's φ is a function of pixel values, not of region size; EP 1 017 019 uses region size to compute belief, not as a separate input to an enhancement parameter. A patentee will argue non‑obviousness on the ground that the references, individually or combined, never teach decoupling region size from belief to modulate an enhancement. The rebuttal is: (i) claim 1 does not exclude size encoded within the belief value, and EP 1 017 019's belief map is by construction a size‑dependent map; (ii) the 573's own specification frames the size term as an optional refinement ("the sharpening may be further refined by the realization that small areas of human flesh… can tolerate higher levels of sharpening… small faces are sharpened less than large faces"), which is a design choice / obvious optimization of the Gouch+belief‑map combination rather than an inventive contribution; and (iii) the monotonic k(S) function is a routine engineering scalar (0 → 1). Expect a claim‑construction fight over whether "in accordance with … the size" requires an explicit region‑size signal.

Other defensive considerations:

• No teaching away. EP 0 681 268 ("do not sharpen faces") and Gouch ("smoothing skin colours") point the same direction as the 573 — toward reduced enhancement in flesh regions. That is consistent art, which strengthens the obviousness case rather than weakens it.
• Secondary considerations. Nothing in the record shows unexpected results, nexus‑bearing commercial success, or long‑felt unmet need beyond the needs the Background itself concedes were already recognized. The prosecution history (family: EP 1 317 475 B1, DE 60203871 T2) and the EPO's treatment of Singhal suggest the belief‑map concept was treated as known.
• PTAB history — do not over‑read it. The page records IPR2019‑01020 (petitioner Unified Patents) as "Not Instituted – Merits." https://portal.unifiedpatents.com/ptab/case/IPR2019-01020 Non‑institution is a § 314(a)/(b) threshold determination, not a merits holding of validity, and it is not evidence of non‑obviousness; conversely it is not evidence of anything either way about the art combinations actually pleaded. Any new analysis should start from the art cited on this page rather than from that docket entry.
• Litigation context. The family shows several E.D. Tex. filings (e.g., 2:23‑cv‑00127, 2:23‑cv‑00169, 5:24‑cv‑00011) and the patent is Expired – Lifetime (expiration adjusted 2023‑07‑13), now held by Monument Peak Ventures. Validity may still matter for past damages, so an obviousness position remains commercially relevant.

6. Bottom line

On the art cited on this page, claims 1, 2, 7, 8, 9, 13, 23–28, 32 and 33 appear vulnerable to a § 103 challenge built on Gouch (US 5,682,443) in view of EP 0 681 268 A1 and EP 1 017 019 A2 (US 6,282,317), further in view of Singhal et al. (Fusion 2000) and Cho/Saber/Lee/Luo‑1996 for the specific enhancement and classification primitives. The motivation is supplied by (a) Gouch's own invitation to automate its weighting function by a "predetermined rule set," (b) EP 0 681 268's express object‑specific sharpening rule, (c) EP 1 017 019's region size/location saliency features producing a graded belief map, and (d) the 573 specification's own admissions about the state of the art. The contestable element is the coupling of region size to a pixel‑wise enhancement control signal; that coupling is arguably the patent's only point of novelty and should be the focus of both sides' claim‑construction and motivation arguments. References cited only under § 102(e) (the Kodak‑owned US 6,738,494; US 6,636,645; US 6,845,181; US 6,891,977; US 2003/0108245 A1; US 2002/0093670 A1) should be treated as corroborating only, given the pre‑AIA § 103(c) common‑ownership defense.