US 6121969

I'll search for this specific patent number and any related CAFC 2026 litigation activity.

Now let me check for any CAFC litigation or PTAB activity involving this patent.

Let me do a couple more targeted checks for litigation and current legal status.

US Patent 6,121,969 — Analyst Summary

1. Bibliographic data

Sources: https://patents.google.com/patent/US6121969/en ; https://uspto.report/patent/grant/6,121,969 ; https://www.freepatentsonline.com/6121969.html ; http://www.everypatent.com/comp/pat6121969.html

2. Abstract (verbatim)

"A similarity-based database of images, where images are preferably ranked and correlated in correspondence to biological preattentive similarity, supports a new type of interface for visual navigation within the database to the end that a human may perceive not only selected images resultant from a query, but the relationship between the selected images. In particular, navigation is within a display space whose geometric characteristics depend on the geometry of the perceptual space in which image similarity is measured. The display space is a subset of the three dimensional Euclidean space that, for many of the distance functions appropriate to the images, is contained in the unit cube. The perceptual intuition of the metric is given (i) in part by the distribution of images in the space, and (ii) in part by making the motion of the user uniform with respect to the metric of the space."

3. Overview of the independent claims (plain language)

The patent has 14 claims. The independent claims appear to be 1, 3, 5, 6, 9, 10, 13 and 14 (claims 2, 4, 7, 8, 11, 12 are dependent).

• Claim 1 — Image data presentation system. A system with (a) an image data store, (b) an input device for querying the images, (c) a computer that ranks stored images by preattentive similarity, giving the images a metric structure that forms an N-dimensional Riemann manifold (dimensions A, B, C … L, M, N) preserving image relationships, and (d) a display that shows the ranked images in a three-dimensional perceptual space (axes X, Y, Z) corresponding to some arbitrary three of the manifold's dimensions, so that relationships among images are made visible.
• Claim 3 — Image database system. A query-responsive image database where conventional matching is replaced by more general similarity measures based on "simple sensorial impressions" in a three-dimensional perceptual space endowed with a Riemann metric more general than ordinary Euclidean distance. The user senses not just the images but their inter-relationships across three displayed dimensions.
• Claim 5 — Man-machine interfacing method (navigation). A method of classifying an image database by an N-dimensional metric, each dimension defined by the user, displaying a subset in a 3-D Euclidean space contained in a unit cube (axes = arbitrary user-selected dimensions), populating that space with images positioned by their metrics, and letting the user navigate by mouse movement, so the user learns both the image relationships and the suitability of the chosen metrics.
• Claim 6 — Recursive man-machine interface method. Manually select three criteria; machine searches/orders the records by them; machine displays a 3-D plot with each criterion as an axis and the reference record at the origin; the human observes clustering/dispersion to formulate new criteria; steps are recursively repeated until a final three-criteria display reveals the records closest to the reference. The point is that the best criteria are typically not known at the outset.
• Claim 9 — Method of displaying object records. Select several criteria, search/order, plot records on the criteria axes with an arbitrary record at the origin, and concurrently display at least some plotted records (visually perceptible and positionally associated with the plot). The combination lets a human judge "at a glance" whether records near/far from the reference are coherently related — i.e., the display is itself a test of the rationality/quality of the ordering criteria.
• Claim 10 — Computer method of displaying digital images. Order images in three human-perceptible criteria A, B, C; define a 3-D display space whose axes are A, B, C and which has a non-linear metric on at least one axis (monotonic in rank order, but equal criterion differences map to unequal axis distances); generate thumbnails; display the thumbnails in the 3-D space per their ordering.
• Claim 13 — Navigable 3-D virtual-reality space. Generating a navigable 3-D space in which separate detached images are ranked and located by three perceptible criteria forming the axes; the viewer can navigate among them, and successive viewer inputs re-rank, re-locate and re-display the images under new viewer-specified criteria, producing successive new 3-D spaces, so the viewer can converge on criteria that isolate images of interest.
• Claim 14 — Iterative man-computer search method. Manually input three criteria; computer creates a 3-D display space with those three metrics and shows thumbnails of the images; the man navigates by mouse movement, noting which metrics discriminate well vs. poorly; repeat iteratively with new criteria derived from experience, until criteria allow finding sought-after images in a database unindexed by any useful criteria.

4. CAFC 2026 docket / litigation check

No CAFC 2026 docket (or any litigation/PTAB proceeding) involving US 6,121,969 was found. I searched for Federal Circuit 2026 argument calendars, appeals referencing the patent number, and general litigation references. The only hits that literally matched the string "6121969" were coincidental artifacts, not this patent:

• A PTAB exhibit file history for IPR2020-01419 (relating to US 9,696,868) in which "6121969" appears only as part of an examiner's prior-art search-string log (("20010050687" | ... | "6121969" | ...)), not as a patent at issue.
• A similar search-string artifact in IPR2020-01417 (relating to US 8,881,048).

Consistent with this, the patent's Google Patents record shows "Expired – Lifetime" with an anticipated expiration of July 29, 2017, which makes active 2026 appellate litigation on this patent unlikely.

5. Uncertainty / caveats

• Claim text provenance. The authoritative full text I was given is truncated before the claims, so the claim summaries above are based on third-party reproductions (uspto.report, FreePatentsOnline, everypatent.com). These contain clear OCR artifacts — e.g., claim 1 renders "arising" as "rising," "manifold" as "manifold," mixes dimension labels ("B, E and H" in claim 1 vs. "B, D, G" in claim 5), and claim 9's text is visibly missing a step phrase. I have not auto-corrected these; treat the exact claim wording as requiring verification against the USPTO/Google Patents PDF (http://patentimages.storage.googleapis.com/1d/78/8b/608fb64281c7c4/US6121969.pdf).
• Number of independent claims. My identification of independent claims 1, 3, 5, 6, 9, 10, 13, 14 is inferred from the reproduced claim set; I could not independently verify claim dependency against the official claim listing.
• Title anomaly. One search snippet from the patent's first-page PDF shows a heading "Image Database Data Navigator" adjacent to "6,121,969." This appears to be an on-screen UI/figure label rather than the patent title (the official title is "Visual navigation in perceptual databases"), but I flag it in case it reflects a drawing-sheet caption.
• Litigation search was not exhaustive. My searches were cut off by a tool limit before I could exhaustively query Docket Alarm, CourtListener, or the CAFC docket system for every 2026 case. The negative result should be read as "no evidence found," not as proof of no proceedings.

I'll research litigation involving US patent 6121969 across patent litigation databases.

Initial results show only unrelated hits (patent search histories, foreign business registrations). Let me search more targeted litigation sources.

No litigation found so far. Let me check dedicated litigation databases directly.

Litigation Search Results for U.S. Patent No. 6,121,969

Bottom line

I found no evidence of any litigation involving U.S. Patent No. 6,121,969. Based on my searches of patent-litigation-oriented sources, this patent does not appear to have been asserted in any federal district court action, ITC investigation, PTAB proceeding (as the challenged patent), or CAFC appeal.

What the patent is (for identification)

To confirm I was looking at the correct patent and not a similar number:

• Patent: U.S. 6,121,969 — "Visual navigation in perceptual databases"
• Inventors: Ramesh Jain (San Diego, CA) and Simone Santini (La Jolla, CA)
• Assignee: The Regents of the University of California (Oakland, CA)
• Appl. No.: 08/902,493 | Filed: July 29, 1997 | Granted: September 19, 2000
• Status: Expired – Lifetime (anticipated expiration July 29, 2017)
• Claim types: Image data presentation system; image database system; man-machine interface method (claims 1–14)

Search results in detail

Every hit associating the string "6121969" with legal proceedings was a false positive:

The PTAB proceedings that surfaced (IPR2020-01417 and IPR2020-01419) involve different patents entirely (8,881,048 and 9,696,868); the '969 number appears in them only as a prior-art search term, not as the subject patent.

Per-case requested fields

Because no cases were identified, there are no plaintiff(s), defendant(s), jurisdiction, case number, filing date, or outcome/status to report for any action involving this patent.

Important caveats on this conclusion

1. Source coverage was incomplete. I was unable to fully load the Unified Patents litigation portal case list (portal.unifiedpatents.com/litigation/caselist) or run the Docket Alarm / PACER-specific queries I had queued before hitting a tool-use limit. My conclusion rests on the searches that did complete.
2. Negative findings are inherently provisional. A patent that is expired (since 2017) and owned by a university is a low-probability litigation candidate, which is consistent with finding nothing — but it does not prove no case ever existed. University-owned patents are occasionally asserted, sometimes years later, and small cases can be missed by general web search.
3. No auto-correction applied. I treated "6121969" literally as U.S. Patent No. 6,121,969 and excluded hits for other numbers (e.g., 6,122,593; 6,121,923; 9,696,868; 8,881,048) even where they appeared in the same documents.

Recommended verification steps (not yet completed)

To convert this provisional "no litigation found" into a defensible negative, the following should be checked directly:

• Unified Patents litigation portal: https://portal.unifiedpatents.com/litigation/caselist (search "6121969")
• Docket Alarm docket search for the patent number as an asserted patent
• PACER Case Locator (search by patent number via the "Nature of Suit 830" filter)
• CAFC docket search for "6,121,969" as the patent on appeal
• USPTO Patent Trial & Appeal Board (PTAB) search — note that because U.S. patents can only be challenged at the PTAB in IPRs/PGRs, and '969 issued in 2000 (pre-AIA, too old for PGR/CBM), only an IPR would be possible, and I found none naming it as the challenged patent.

If you'd like, I can attempt these targeted database queries one at a time so we can reach a firmer conclusion.

I'll verify the empty structured PTAB list against public sources before writing anything up.

Let me run two more targeted checks — the PTAB E2E / FWD search and any ex parte reexam (non-AIA but relevant) before concluding.

PTAB / AIA Trial Proceedings — US Patent 6,121,969 ("Visual navigation in perceptual databases")

Cross-reference note: This builds on the earlier bibliographic, claims, litigation, and PTAB sections. The structured block above ("PTAB proceedings on file") states that the USPTO ODP API returns no AIA trial proceedings for this patent. I attempted to falsify that negative with targeted web searches (PTAB E2E / FWD indices, Docket Alarm, and docket-scraped PTAB documents). The searches confirm the empty list — see the false-positive analysis below. No contradiction with prior sections.

Proceedings overview

Total AIA trial proceedings on US 6,121,969: 0. Breakdown by status: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution-denied — the patent has never been the subject patent in any Inter Partes Review, Post-Grant Review, or Covered Business Method Review.

Bottom line for a defendant: there is no PTAB record to lean on, and there never will be one arising from past filings. The '969 patent expired on 2017-07-29 and its term is over, so the defensive posture is not "the patent is hardened by surviving IPRs" nor "claims 1–5 are canceled." It is simpler and stronger: the patent is expired — no claim can be infringed on or after its expiration, and no petitioner has ever put a claim at risk. Conversely, if a demand letter is circulating in 2026 that cites '969, the letter is almost certainly relying on past (pre-2017) conduct, and the absence of any PTAB challenge means there is no free-standing PTAB invalidity record you can point to.

Caveat on the "zero" finding (stated explicitly, per the operative rules): a negative is provisional. But here the negative is strongly corroborated by two independent facts: (a) the structured ODP list is empty, and (b) the patent expired in 2017 and is university-owned, a low-probability assertion/IPR profile. I could not locate even a dismissed or denied petition, which for a patent this old is consistent with "never challenged."

Individual proceedings

None to report. There is no proceeding number, petitioner, panel, institution decision, Final Written Decision, settlement, or appeal to document for US 6,121,969. I will not invent a proceeding number to fill this section.

False positives I ruled out (so the record is clean)

Because the number is easy to confuse, here is every "6,121,969 / 6121969" hit I found in PTAB-adjacent sources, and why each is not a proceeding on this patent:

No hit showed '969 as the challenged patent — i.e., no "Petitioner v. Regents of the University of California" caption, no IPR####-##### caption naming the '969 patent as the subject.

Strategic summary

Claim status. Because no AIA trial ever reached a Final Written Decision on this patent, no claim of US 6,121,969 has been canceled, narrowed, or held unpatentable by the PTAB. All 14 claims stand as issued — but this is largely academic, because the patent's statutory term expired 2017-07-29. "Untested" is the accurate label for every claim (1–14): not canceled, not sustained, never adjudicated. A defendant should therefore not expect to find an FWD to cite; there is none.

Estoppel landscape. § 315(e)(2) estoppel is irrelevant here — estoppel only attaches to a petitioner who instituted an IPR, and there is no such petitioner. That cuts both ways: (i) there is no petitioner-side estoppel constraining anyone, and (ii) there is no institution to have triggered it. Practically, because the patent is expired, the more relevant doctrines are § 286's six-year damages look-back and the general rule that expiration ends the accrual of new infringement claims. If a dispute exists at all, it is almost certainly a stale/expired-patent demand, which should be met with a § 286 limitations and laches/estoppel analysis rather than a PTAB strategy.

Pattern signals. No pattern exists: no repeat petitioner, no serial IPRs, no joinder, no PTAB appeal, no defensive aggregator (Unified Patents and similar NPE-focused filers never touched this patent). PGR and CBM were never available to any challenger: '969 issued 2000-09-19 from a pre-AIA application (filed 1997-07-29), so it is outside PGR (post-AIA filings only) and, as a 2000-issued patent that predates the CBM transitional program's business-method scope, an unlikely CBM target — the CBM window also closed 2020-09-16. That leaves only IPR as a theoretical vehicle, and none was ever filed.

Why the absence is itself informative. The prior sections note the patent was never asserted in district court. Combine that with zero IPRs and you get a coherent story: this was an academic/university patent that was never commercially enforced. Well-asserted patents reliably attract IPRs; a patent that was never litigated never needed one. So the absence of PTAB activity here is a neutral-to-favorable signal for a defendant, not a warning that the patent is untouchable.

Recommended next steps

• Do not plan around a PTAB record — there isn't one. Any defense built on "claims X were canceled in IPR" is unavailable by definition. If a demand letter cites '969, respond on expiration (2017-07-29) and § 286 damages limitation grounds, not on PTAB history.
• Verify the expired status independently before relying on it: Google Patents legal-status field shows "Expired – Lifetime," anticipated expiration 2017-07-29 (https://patents.google.com/patent/US6121969/en), sourcing USPTO maintenance-fee records. Confirm against USPTO Patent Center (https://patentcenter.uspto.gov, app. 08/902,493) that no maintenance-fee reinstatement extended the term — a reinstated patent would change the analysis.
• Run the definitive PTAB negative yourself if you need it for a filing: PTAB E2E search (https://ptacts.uspto.gov/ptabweb) for patent number 6121969, and the PTAB Decisions page (https://www.uspto.gov/patents/patent-trial-and-appeal-board/ptab-decisions). Docket Alarm returned only the search-string artifacts above; CourtListener (https://www.courtlistener.com) shows no CAFC appeals of any '969 FWD because no FWD exists.
• Non-AIA check (optional but cheap). The ODP block covers AIA trials only. If due diligence demands completeness, confirm there is no ex parte reexamination (which is not an AIA trial and would not appear in the structured list) via the Patent Center reexam tab for 08/902,493 — though given expiration, even a reexam would have no live effect.
• Bottom line: plain and simple — no PTAB activity exists on US 6,121,969, in any status, ever. The patent is expired and was never challenged at the Board.

I'll research the assignment record for US 6,121,969, starting with the USPTO assignment databases and cross-referencing legal events.

Found the primary assignment record. Let me verify there are no subsequent transfers and identify the correspondent.

I hit the tool-call limit while verifying the absence of post-issuance transfers. Here is the assignment analysis based on the records retrieved.

Assignment Record — US Patent 6,121,969 ("Visual navigation in perceptual databases")

Cross-reference note: This builds on the earlier bibliographic, litigation, and PTAB sections. It is consistent with them: the litigation section found no district-court assertion and the PTAB section found zero AIA trials. A single, never-reassigned university assignment is exactly the ownership profile that predicts that outcome. No contradictions found.

Primary source caveat: The USPTO Assignment Center search interface (https://assignmentcenter.uspto.gov/) was not directly loadable in this session. The assignment below is drawn from the patent's own front-page/legal-events data as republished by Google Patents (https://patents.google.com/patent/US6121969/en), which mirrors the Assignment Center record and gives the reel/frame. Treat the reel/frame as verified by one authoritative mirror, not yet by the USPTO raw record.

Inventors

Pattern check — unusual departures: Not present. Both inventors were in the same UCSD lab (the "Visual Computing Laboratory" run by Jain) at filing, and both assigned to their employer of record in the ordinary course. There is no evidence of inventors departing the original assignee within 12 months of filing, and no portfolio fire-sale followed. Jain's later move to UC Irvine is an individual academic move, not a transfer of this patent. The self-citations (Santini 96a/96b/96c, 97a) confirm the work was done inside the UCSD group — an organic university-originated invention, not an acquired-in patent.

Original assignee

The Regents of the University of California (Oakland, CA); original assignee field on the issued patent reads "University of California San Diego UCSD." Assignor-side filing assignment names "REGENTS OF THE UNIVERSITY OF CALIFORNIA, THE, CALI[FORNIA]."

• Primary line of business: public university / university research administration (a land-grant public institution; UC San Diego is one of its campuses). UC does not sell products from this patent.
• Product embodying the claims: No. This is a research/experimental interface — a "visual navigator" for content-based image retrieval demonstrated in a lab setting. There is no evidence UC commercialized or shipped any product embodying claims 1–14. The patent is a research disclosure captured by the university's technology-transfer office, not a product patent.
• Current status: Operating. The University of California is a going concern and the current recorded owner. Its intellectual-property holding entity is the Regents. The patent itself is Expired – Lifetime; term ended 2017-07-29 for failure to keep it in force beyond the 12-year maintenance fee window.

Assignment timeline

• 1997-07-17 to 1997-07-19 (executed) / recorded 1997-07-29 — Reel 008735/0210

  • Conveyance: Assignment (ASSIGNMENT OF ASSIGNORS INTEREST)
  • Assignor: Jain, Ramesh; Santini, Simone
  • Assignee: Regents of the University of California, The (California)
  • Correspondent: Not stated in the data retrieved. (Note: the prosecution attorney of record on the issued patent is the firm Fuess & Davidenas, La Jolla, CA; the assignment-record "correspondent" field — the filer of the recording — was not retrieved and I will not guess it. Per the operative rules I flag this as an open field rather than fabricate a name.)
  • Context: Original filing assignment — inventors convey their rights to their employer, the University of California, in the ordinary course of employment (a standard university-style obligation-to-assign). This is the base-case "internal/employment" assignment, not a sale.

• No further assignment events recorded. The Google Patents legal-events log for this patent contains only:

  • 1997-07-29 — AS (Assignment) — the single event above (REEL/FRAME:008735/0210).
  • 2000-09-19 — STCF (Patent grant) — issue event, not a transfer.
  • FPAY × 3 — maintenance-fee payments at years 4, 8, and 12 — fee events, not transfers.
  • FEPP — status change from small entity to large entity / undiscounted (UC stopped claiming small-entity status). A fee-status change only, not an ownership change.
  • 2017-07-29 — Anticipated expiration / Expired – Lifetime.

There is no recorded post-issuance assignment, no security agreement, no merger, no change of name, and no license recording. The chain is exactly one link long: inventors → Regents of the University of California, and it has stayed there since 1997. That is itself the finding.

Timeline diagram

timeline
    title Ownership of US 6121969
    1997 : Inventors execute assignment Jul 17-19
         : Recorded reel 008735 frame 0210
         : Application filed Jul 29
    2000 : Patent issued Sep 19
    2003 : Maintenance fee year 4 paid
    2007 : Maintenance fee year 8 paid
    2011 : Maintenance fee year 12 paid
    2012 : Entity status changed to large entity
    2017 : Patent expired Jul 29

NPE / troll-pattern signals

Verdict

Insufficient data — and specifically, only the original assignment is on record.

The complete ownership chain is a single recorded conveyance: inventors Ramesh Jain and Simone Santini → The Regents of the University of California, executed 1997-07-17 to 1997-07-19 and recorded 1997-07-29 at Reel 008735/Frame 0210, with no post-issuance assignment of any kind before the patent expired 2017-07-29. This is the textbook "no NPE activity" profile: a university-employed-inventor assignment that was never sold, never securitized, never licensed of record, and never asserted — so there is no shell entity, no known asserter, no repeat NPE correspondent, and no cascading chain to evaluate. None of the eight NPE signals can even be evaluated on a single-link chain, which is why the verdict is "insufficient data" rather than an affirmative "operating-company assertion" call.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent number 6121969; the mirror at https://assignment.uspto.gov/patent/index.html surfaces the same record). Confirm the reel/frame 008735/0210 and confirm that no later AS entry exists, and pull the assignment-record correspondent field that I could not retrieve.

Two open items I could not close before the tool limit (stated explicitly, not fabricated):

1. Assignment-record correspondent (the filer of the 1997 recording) — not retrieved. Flagged for direct lookup.
2. Exhaustiveness of the "no later assignments" negative — I confirmed the single event via the Google Patents legal-events mirror, which is authoritative-for-practical-purposes but is not the USPTO raw record; the direct Assignment Center query should be run to make the negative airtight.

I'll research the prior-art references cited on the face of US 6,121,969 and in its specification. Let me pull details on each.

Let me research the remaining cited references.

Prior-Art Analysis — U.S. Patent No. 6,121,969 ("Visual navigation in perceptual databases")

Scope note / method. I treated the identifier "6121969" literally as US 6,121,969 and excluded similar numbers (e.g., 6,122,593, 9,696,868, 8,881,048). Because the patent was filed 1997‑07‑29 (pre‑AIA), the pre‑AIA version of 35 U.S.C. § 102 governs. Two threshold dates matter:

• § 102(b) critical date = 1996‑07‑29 (one year before filing).
• § 102(e) = a U.S. patent/application filed before the '969 filing date, even if it issued after it.

I built this from (i) the authoritative full text you supplied, which lists the face-of-patent references, and (ii) web verification of the individual references. I was cut off by a tool-use limit before I could verify every bibliographic detail, so items I could not confirm are flagged explicitly rather than filled in.

Important framing caveat: § 102 anticipation requires a single reference to disclose every limitation of a claim as arranged. On the record available, none of the cited references alone appears to disclose the '969's core novelty (a Riemann/perceptual-space metric with 3‑D visual navigation). The references are therefore best characterized as § 103 prior-art fodder, and I say so below rather than overstating § 102. Where a reference maps to specific claims, I identify them.

1. Face-of-patent U.S. patent references

Per the earlier bibliographic section, the following U.S. patents are cited on the face of '969: US 5,579,471 (Barber); 5,636,350 (Eick); 5,751,286 (Barber); 5,847,708 (Wolff); 5,886,698 (Sciammarella); 5,893,095 (Jain et al.); plus the Campanai et al. NPL item.

1a. US 5,579,471 — Barber et al., "Image Query System and Method" (IBM) — VERIFIED

• Citation / dates: U.S. Patent 5,579,471, issued 1996‑11‑26; App. No. 216,986, filed 1994‑03‑23; continuation‑in‑part of Ser. No. 973,474 filed 1992‑11‑09 (abandoned). Assignee International Business Machines Corp. Inventors: Barber, Beitel, Equitz, Niblack, Petkovic, Work, Yanker.
• § 102 category: § 102(a) (patented before the '969 invention/filing) and § 102(e) (filed 1994, before 1997‑07‑29). Not § 102(b), since issuance was only ~8 months before the '969 filing.
• Description (verified from the patent PDF and uspto.report): An image database searched by visual characteristics — color, texture, shape, size — as well as textual tags. Queries are composed in an image query construction window by dragging thumbnails (sample image-characteristic icons) into position; a query engine returns database images whose features match the placed thumbnails. This is the canonical "query-by-image-content" (QBIC-family) reference.
• Claims potentially implicated: The image-store + query-input + computer + image display substrate of claim 1; the thumbnail element of claim 10; the content-based similarity search underlying claims 3 and 14.
• § 102 assessment: No full anticipation of any independent claim. Barber discloses content-based image query and thumbnail-based UI, but nothing about a Riemann manifold, a unit-cube 3‑D perceptual/display space, or metric-driven navigation. Best used as primary art in a § 103 combination.

1b. US 5,751,286 — Barber et al., "Image Query System and Method" (IBM) — VERIFIED

• Citation / dates: U.S. Patent 5,751,286, issued 1998‑05‑12; App. No. 788,150, filed 1997‑01‑24; a continuation of Ser. No. 720,993 (filed 1996‑10‑15, abandoned), which continued Ser. No. 216,986 (now US 5,579,471). Assignee IBM.
• § 102 category: § 102(e) only — it issued (1998) after the '969 filing, but its application was filed (1997‑01‑24) before 1997‑07‑29.
• Description: Substantively the same disclosure as '471 (same family/spec): content-based image search by color/texture/shape/size plus text tags, with the drag-and-drop query-construction interface and thumbnails.
• Claims potentially implicated: Same as '471 — claims 1, 3, 10, 14.
• § 102 assessment: Same as '471 — partial, not anticipatory. Because it is the same family, it should not be treated as cumulative additional art beyond '471.

1c. US 5,636,350 — Eick et al., "Using Symbols Whose Appearance Varies to Show Characteristics of a Result of a Query" (AT&T) — VERIFIED

• Citation / dates: U.S. Patent 5,636,350, issued 1997‑06‑03; a continuation of Ser. No. 08/111,087, filed 1993‑08‑24. Assignee AT&T / Lucent.
• § 102 category: § 102(a) (patented 1997‑06‑03, before the '969 filing) and § 102(e) (parent filed 1993‑08‑24).
• Description (verified from patent PDF): A graphical presentation of database query results. A grid is displayed in which each cell represents a query result, and the appearance of a symbol (e.g., color, size) varies with properties of the hits (number of hits, pages). Bar graphs along the edges show totals; sliders filter thresholds; moving a pointer over a symbol and clicking surfaces the underlying references. Explicitly oriented to "graphical displays of the results of queries in particular" and to rapid visual interpretation of search output.
• Claims potentially implicated: The display-the-results-of-searching-and-ordering step and the visual/concurrent display elements of claim 9; the criteria-as-axes visual plotting concept of claim 6; the interactive re-display of claim 13.
• § 102 assessment: Strongest § 102(candidate) on the results-visualization claims, but still not a full anticipation of claims 6/9/13 — Eick shows 2‑D grid/bar-chart visualization of query results, not a 3‑D plot with one criterion per axis, reference record at the origin, and thumbnails positioned in the plot. Very useful as § 103 art against the "display results in a plot so a human can judge quality" idea.

1d. US 5,847,708 (Wolff) — NOT VERIFIED

• I could not independently confirm the title, issue date, filing date, or subject matter within the tool budget; my targeted query returned unrelated key-frame/video material, which is a false positive. The face of '969 lists Wolff as the inventor. Treat the following as provisional: if it issued after 1997‑07‑29, it qualifies only as § 102(e) art (requiring a pre‑1997‑07‑29 filing date). Do not rely on this entry without pulling the document itself (I recommend fetching patents.google.com/patent/US5847708).

1e. US 5,886,698 (Sciammarella) — NOT VERIFIED

• Similarly unverified; the face of '969 lists Sciammarella. Issue date appears to be 1999 (post‑filing), which would make it § 102(e) art at best, contingent on a pre‑1997‑07‑29 filing date. Verify before use.

1f. US 5,893,095 — Jain et al. — PARTIALLY VERIFIED

• Verified facts: A U.S. patent numbered 5,893,095, issued 1999‑04‑06, inventor Jain et al. (confirmed from a citation listing). Same-inventor/same-assignee (UCSD) lineage is likely.
• Unverified: exact title and filing date. It is plausibly a sibling application filed around the same time as '969 (App. 08/902,493).
• § 102 analysis — important nuance: If '095 was filed on or after 1997‑07‑29, it is not § 102(e) art, because § 102(e) requires the reference application to be filed before the applicant's filing/invention. A same-day or later filing therefore does not anticipate '969. If it was filed before 1997‑07‑29, it is § 102(e) art. This determination turns entirely on the filing date, which I could not confirm.
• Claims potentially implicated: Broadly the similarity-based image-database subject matter of claims 1 and 3.

2. Prior art discussed in the '969 specification (Background, § 2.2)

Unlike the face references, these four patents are discussed in the text and are all firmly § 102(b) art (all issued before 1996‑07‑29):

2a. US 4,899,290 — Hartzband et al., "System for specifying and executing protocols for using iterative analogy and comparative induction in a model-based computation system" (Digital Equipment Corp.)

• Issued 1990‑02‑06 (filed earlier). § 102(b).
• Description (from the '969 text): Performs iterative specialization and generalization among objects in a set; performs an analogy / symmetric comparison between a pair of objects to produce similarity and difference reference structures, then iteratively repeats comparison using the previously generated reference structure.
• Claims implicated: The recursive/iterative criteria-development claims — claim 6 and claim 14 — and arguably claim 9.
• § 102 assessment: The closest art on iterative similarity refinement, but Hartzband operates on symbolic/model-based objects, not on a 3‑D geometric display whose axes are the criteria, and does not show human visual observation of clustering/dispersion to reformulate criteria. Not a clean anticipation; strong § 103 material for the "recursive" limitation.

2b. US 5,267,329 — Ulich et al., "Process for automatically detecting and locating a target from a plurality of two dimensional images" (Kaman Aerospace Corp.)

• Issued 1993‑11‑30. § 102(b).
• Description: Machine (fully automatic) processing of 2‑D images — noise reduction, spatial filtering, asymmetric threshold detection, contrast stretching, localization, recognition — to detect/locate a target with high probability of detection and low false-alarm rate; optionally displays an output image for operator visual confirmation.
• Claims implicated: Tangentially the image-classification/ordering notions of the independent claims.
• § 102 assessment: Not anticipatory. The '969 spec itself distinguishes it: classification here is entirely by machine, whereas '969 requires a human-navigable perceptual space.

2c. US 5,325,466 — Kornacker, "System for extracting knowledge of typicality and exceptionality from a database of case records" (Perceptive Decision Systems, Inc.)

• Issued 1994‑06‑28. § 102(b).
• Description: A knowledge-tree system that iteratively partitions a database of case records into conceptually meaningful clusters, each automatically assigned a meaning by its pattern of typicality/exceptionality, with no prior domain knowledge required. Suited to pattern classification/nonmonotonic reasoning.
• Claims implicated: The clustering/observing-distribution / formulate-new-criteria steps of claims 6 and 9.
• § 102 assessment: Relevant to the cluster-recognition limitation, but the '969's clustering is a visual, human-observed dispersion in a 3‑D display, not an automated tree partition. Again § 103, not anticipation.

2d. US 5,416,892 — Loken-Kim & Kyung-ho, "Best first search considering difference between scores" (Fujitsu Ltd.)

• Issued 1995‑05‑16. § 102(b).
• Description: A best-first search for AI problem-solving using a novel search priority index computed from the difference between scores of a node and the next node in breadth, reducing the number of search steps to a solution.
• Claims implicated: The search/order steps of claims 6, 9, 14.
• § 102 assessment: Concerns search-algorithm efficiency, not a visual perceptual-space interface. Not anticipatory; marginal § 103 art.

3. Non-patent literature cited on the face

3a. Campanai et al., "Using 3D Spatial Relationships for Image Retrieval by Contents," 1992 IEEE Workshop on Visual Language Proceedings, pp. 184–190 — VERIFIED as cited

• Publication date: 1992 → § 102(b) (published >1 year before 1997‑07‑29).
• Description (from title/citation): Image retrieval that models/uses 3‑D spatial relationships as the retrieval key.
• Claims implicated: The 3‑D display / spatial-relationship aspects of claims 3, 5, 10, 13.
• § 102 assessment: The closest NPL to the "3‑D spatial display of images" idea, but it is a retrieval-representation paper, not a navigable unit-cube perceptual-space interface. Useful § 103 art.

3b. The inventor's own papers (NOT prior art — flagged for completeness)

The spec cites Santini 95, Santini 96a, Santini 96b, Santini 96c, Santini 97a (Santini & Jain). The '969 text expressly states these are the inventors' own work "and are thus not prior art." They should be excluded from any § 102/§ 103 rejection absent a § 102(b)/§ 103(a) independent basis. (Also cited: Tversky 77, Treisman 86/87, Shepard 87, Pentland 94, Flickner 95, etc. — scientific/technical background, not claim-level art.)

4. Bottom-line § 102 assessment

Conclusion. No cited reference, alone, discloses the '969's distinguishing combination: (i) an N‑dimensional Riemann/perceptual metric built from preattentive-similarity features, (ii) a unit-cube 3‑D display space whose axes are three selected similarity criteria, and (iii) metric-adaptive navigation in which viewer motion is uniform with respect to the local metric. The most relevant art clusters as:

• Content-based image query / thumbnails: Barber '471 and '286.
• Visualization of query results / interactive result display: Eick '350.
• Iterative/recursive similarity refinement and clustering: Hartzband '290 and Kornacker '466.
• 3‑D spatial representation for image retrieval: Campanai et al.
• Search efficiency: Loken-Kim '892.

These are properly deployed as a § 103 combination (Barber or Campanai as primary, Eick and Hartzband/Kornacker as secondary), not as § 102 anticipations.

5. Explicit caveats (per the “do not fabricate” rule)

1. Unverified entries: US 5,847,708 (Wolff) and US 5,886,698 (Sciammarella) — I could not confirm title, dates, or content; the one search returned a false-positive on unrelated key-frame video art. Do not rely on these rows until the documents are pulled.
2. US 5,893,095: issue date (1999‑04‑06) and inventor (Jain et al.) are verified; title and filing date are not. The § 102(e) status hinges entirely on whether it was filed before 1997‑07‑29.
3. Claim numbering (independents 1, 3, 5, 6, 9, 10, 13, 14) is carried over from the earlier generated section and rests on third-party reproductions; the authoritative text you supplied is truncated before the claims. Verify against the granted PDF: patentimages.storage.googleapis.com/1d/78/8b/608fb64281c7c4/US6121969.pdf.
4. No auto-correction applied: all identifiers are read literally; I excluded look-alike numbers (6,122,593; 9,696,868; 8,881,048) even where they co-occurred in search hits.
5. No source yet found lists these references as formal § 102 rejections — I analyzed them as candidate art. A full file-wrapper review (USPTO Patent Center, App. 08/902,493) would confirm which were actually applied by the examiner and on what statutory basis.

I'll ground this analysis in the actual prior-art references before reasoning about combinations. Let me pull the key cited references.

Let me pull the remaining cited references.

Obviousness Analysis — U.S. Patent 6,121,969 under 35 U.S.C. § 103

Cross-reference note. This builds on the earlier sections (bibliographic data, claim summaries, litigation, PTAB). I do not repeat them. Two earlier caveats are load-bearing here and are restated because they bound the reliability of this opinion: (i) the authoritative full text I was given is truncated before the claims, so claim element recitations below are taken from third-party reproductions flagged in the summary section; (ii) the "Number of independent claims" identification was inferred. Where a limitation is quoted below, treat it as requiring verification against the issued claim set before it is relied on in a filing.

1. Legal framework and threshold dates

• Governing statute: The application was filed 1997-07-29 and issued 2000-09-19. It is a pre-AIA patent, so pre-AIA § 103(a) applies, with the Graham v. John Deere framework plus the KSR Int'l v. Teleflex (2007) rationales for combining references. KSR applies to all cases regardless of filing date.
• Critical date: 1997-07-29. A reference is available as § 102(b) art only if published more than one year before that date (i.e., before 1996-07-29); as § 102(a)/(e) art if earlier in date (issuance or filing).
• Scope of this opinion: The patent expired 2017-07-29 (per the earlier sections). This analysis is therefore directed at retrospective validity (e.g., a stale demand letter, a § 286 look-back dispute, or a defensive posture), not at a live infringement defense.

2. Person having ordinary skill in the art (POSITA)

A POSITA here would be a computer scientist or electrical engineer with 2–4 years' experience in content-based image retrieval (CBIR) and database/GUI systems, or a master's degree with equivalent exposure, familiar with: color/texture/shape feature extraction and similarity metrics; multi-dimensional indexing; and interactive data visualization (scatter plots, multi-view/linked displays, 3-D scientific visualization). The patent's own Background confirms the field was mature and crowded by 1997 ("Perceptual databases … have received a lot of attention circa 1997," citing Pentland 94, QBIC/Flickner 95, Jacobs 95, etc.).

3. The prior art actually on the page

The "Prior Art" material divides into two groups: (a) background references cited in § 2.1 of the specification (Pentland 94, Flickner 95, Treisman 86, Malik 90, etc.), and (b) the references on the face of the patent (the list carried over from the earlier section). I verified the substance of three of these directly; the rest I flag.

4. Claim-element decomposition and the two lead combinations

4.1 Combination A — claims 1, 3, 5

Barber '471 (primary) + Eick '350 + Campanai 1992 + Pentland 94/Flickner 95 (state of the art)

Gap: nothing in these references uses the words "Riemann manifold," "metric tensor," or "unit cube." That is the entire non-obviousness case for claims 1 and 3 (see §6).

4.2 Combination B — claims 6, 9, 14

Barber '471 + Eick '350 + Sciammarella '698

Gap: the express recitation that the same three criteria define both the plot axes and the ordering and that iteration converges on the "best" three criteria is a workflow characterization of the routine refine-and-re-query loop that Barber already describes. Under KSR, arranging a known refinement loop into three user-chosen axes on a 3-D plot is a predictable application.

4.3 Combination C — claims 10, 13

Barber '471 + Sciammarella '698 + Eick '350

Claim 10's distinctive limitations are (a) thumbnails and (b) a non-linear metric on at least one axis ("monotonic in rank order" but "equal criterion differences map to unequal axis distances").

• Thumbnails: Barber '471 (thumbnails are the claimed display primitive; QBIC "reduces [images] to a standard-sized icon called a thumbnail").
• Non-linear axis metric: Sciammarella '698 maps relevance to area (non-linear visual encoding); Eick '350 maps hit count to symbol size. Once a POSITA uses relevance as a coordinate, applying a monotonic non-linear scale (log, rank, or relevance-proportional) is a routine design choice. The '969 specification itself describes the intended effect — images concentrating near the origin, sparse at the border — which is exactly a non-linear density mapping.
• Claim 13's successive re-ranking/re-location under new viewer input is Barber '471's refinement/navigation loop instantiated in the Eick/Sciammarella spatial display.

5. Motivation to combine (KSR rationales, applied)

Each rationale is tied to record evidence rather than generic boilerplate:

1. Same field of endeavor and same problem. Barber '471, Eick '350, and Sciammarella '698 all address displaying and refining the results of a database query to a human user. The '969 problem ("it is an entirely unanswered question as to just how, in detail, navigation within the perceptual space should proceed"; "how the query should be modified if the results are not satisfactory") is the same problem.

2. Predictable result / combination of known functionalities. Barber supplies the CBIR engine and the navigation/refinement loop; Eick and Sciammarella supply the spatial, image-based result display with a visual variable. Combining a known retrieval engine with a known visualization of its output is the paradigm "predictable" combination — nothing in '969's display does more than expose the retrieval engine's distances.

3. Known technique / obvious improvement. 3-D scatter-plotting of multi-attribute data and moving a viewpoint through a 3-D scene were standard visualization practice by 1997 — the '969 specification itself concedes the analogy: the user "will never reach a 'border' of the space … as is suggested by the common experience of navigating an avatar in the three-dimensional space of an interactive computer game."

4. Design incentive / necessity of dimension reduction. The human visual system displays three spatial dimensions. Condensing an N-dimensional similarity space to three is not an insight; it is a constraint. The '969 claim language itself covers "some arbitrary three" of the manifold's dimensions (claim 1) and "arbitrary user-selected dimensions" (claim 5) — i.e., it claims the choice of which three, which is the least inventive part of the operation.

5. Explicit teaching in the primary reference toward the combination. Barber '471 does not merely permit refinement — it advocates "query refinement and navigation through the image database" and teaches similarity measures selected to match human perception. That is the strongest available motivation language, and it comes from the primary reference.

6. No teaching away. None of the cited references disparages spatial display of image results or iterative refinement. The only "warning" in the record is the patent's own observation that the correlates "need not, and generally do not, match the relatively homogeneous feature distribution occurring in a virtual reality computer game" — a caveat, not a teaching away.

6. Where the claims are strongest — and why it likely does not save them

The Riemannian/metric-tensor formalism is the entire non-obviousness story for claims 1, 3, 5, 13. Two responses:

• Mere verbal characterization / mathematical description of a known technique. The specification's own theory section defines the "Riemann" distance as an integral of a local metric over a curve joining two points, i.e., a generalized weighted distance between similarity vectors — the same computation QBIC already performed with non-Euclidean, cross-term quadratic distance expressions and weighted Euclidean texture/shape distances. Under KSR ("the combination of familiar elements according to known methods"), relabeling a weighted similarity distance as a "metric tensor" and a feature space as a "Riemann manifold" does not make the underlying data processing patentable. The asserted advance over QBIC reduces to terminology plus a display convention.
• But be candid about the residual risk. A court could accept that claim 5's velocity-adapted navigation ("making the motion of the user uniform with respect to the metric of the space") and claim 13's successive re-ranking into new 3-D spaces are not squarely shown by Barber/Eick/Sciammarella. The metric-adapted speed limitation appears in the abstract and specification, but I could not confirm which claim (if any) actually recites it — the claim text I have is truncated and partly corrupted. If it sits in a dependent claim, § 103 falls back on the independent claims' broader "navigate by mouse movement" language, which Barber and Eick both meet.

Caveat on claim 9: the earlier section records that the reproduced claim 9 text is visibly missing a step phrase. Any § 103 chart for claim 9 is provisional until the issued text is confirmed.

7. Prior-art date and availability problems to verify before relying on this

1. Sciammarella '698 issued 1999-03-23 — after the '969 filing date. It is available against '969 only as § 102(e) art, i.e., only if its own filing predates 1997-07-29. Because it appears on the face of '969, it was presumably cited as § 102(e) art, but confirm the filing date before using it. If its filing postdates '969's, drop Combination B/C's reliance on it and substitute Eick '350 alone (which covers relevance-encoded symbol size but not the image-pyramid display).
2. Barber '471 and Eick '350 are safe on their face (issued 1996-11-26 and 1997-06-03), subject to a routine § 102(a)/(e) date check.
3. US 5,893,095 (Jain et al.) is very likely disqualified. If it is the work of the same inventive entity (Jain and/or Santini), it is not "by another" under § 102(a)/(e), and if it is merely commonly owned it is disqualified for § 103 purposes by pre-AIA § 103(c). Do not build a combination on it without confirming the inventive entity.
4. Santini 96a–c and 97a are admissions of non-prior-art — the patent says so expressly. Do not cite them as art.
5. US 5,847,708 (Wolff) and US 5,751,286 (Barber) were not verified in this pass. I will not assert what they disclose.

8. Bottom line

• Strongest § 103 case: Barber '471 (primary) + Eick '350 + Campanai 1992, optionally with Sciammarella '698 (dated), renders claims 1, 3, 5, 6, 9, 10, 13 and 14 obvious. Every functional element — content-based image store, query by example, similarity ranking in lieu of matching, thumbnails, multi-axis spatial display of results, mouse navigation, iterative criterion refinement — is distributed across these references, and the combination is supported by KSR (same field, same problem, predictable result, known technique, design incentive, and an express motivation in the primary reference itself).
• The one genuinely contested element is the Riemannian/metric-tensor characterization, which is a mathematical restatement of the non-Euclidean similarity distances already taught by QBIC ('471 / Flickner 95). The best attack is "mere verbal characterization of a known technique," not anticipation.
• The claims most likely to survive are any dependent claims that specifically recite velocity adapted to the local metric and the closed/curved, border-less navigation — but I could not confirm which claim(s) recite those, because the claim text I was given is truncated and partly corrupted. Verifying the issued claim set against the Google Patents PDF (http://patentimages.storage.googleapis.com/1d/78/8b/608fb64281c7c4/[US6121969](/patent/US6121969).pdf) is the single most valuable next step before this analysis is used for any purpose.
• Flagged contradiction with the earlier sections: none discovered. This section is consistent with the earlier note that the claim summaries rest on third-party reproductions, and it adds two new qualifications — Sciammarella's § 102(e) date problem and US 5,893,095's same-inventor disqualification — that were not addressed previously.