US 5751232

US5751232 — Concise Patent Overview

Title: High-efficiency encoding apparatus and high-efficiency decoding apparatus
Patent number: US5751232A
Assignee: Mitsubishi Electric Corp (listed as both current and original assignee)
Inventors: Sadayuki Inoue; Junko Ishimoto; Ken Onishi
Priority date: 1993-07-30
Filing date: 1996-08-28 (application US08/704,276)
Publication/grant date: 1998-05-12
Legal status: Expired – Fee Related; anticipated expiration 2014-07-29

Purpose and plain-language scope:
The patent describes an apparatus for encoding data at high efficiency and a corresponding apparatus for decoding it. The supplied excerpt identifies recurring concepts used to characterize the invention: "data," "length," "run," "coefficient," and "variable." Taken literally, this indicates the encoding/decoding scheme deals with runs and coefficients of data, with lengths and variable-length aspects — consistent with an entropy or variable-length coding approach. Beyond these keyword listings, the excerpt does not describe the specific encoding method, algorithm, or claim limitations.

Not established by the supplied excerpt:

• The full text of any claim, including independent claim scope and dependency structure.
• The specific technical steps of encoding or decoding.
• Any cited prior-art references, examiner citations, or prosecution history.
• A detailed description, drawings content, or embodiments (the excerpt contains image placeholders only, no figure content).
• The relationship, if any, to related applications or priority documents beyond the 1993-07-30 priority date shown.
• Ownership transfers, litigation, or validity conclusions — expressly out of scope and not researched.
• Whether the listed "current assignee" is accurate; the source itself states the assignee list may be inaccurate and that legal status is an assumption, not a legal conclusion.

Source: https://patents.google.com/patent/US5751232/en (fetched 2026-09-26). The provided material is an excerpt; the source should be consulted for omitted text, claims, and drawings.

I'll search for litigation involving US patent 5751232. Let me run several targeted searches.

Let me search more specific litigation databases.

Let me try more targeted litigation searches.

Summary of Findings

No litigation involving U.S. Patent No. 5,751,232 was identified in the searches performed.

I searched for litigation records tied specifically to patent number 5,751,232 (both "5751232" and "5,751,232" formats) across sources that index patent litigation, including the Unified Patents portal (which returned only the patent's bibliographic page, not a litigation/case list), general patent-litigation web results, PACER-adjacent aggregators, and CourtListener/Justia materials. No case asserted this patent.

The patent itself (for identification)

Source: Google Patents (https://patents.google.com/patent/US5751232/en) and Unified Patents (https://portal.unifiedpatents.com/patents/patent/US-5751232-A).

Important cautions about the negative result

1. Do not treat the negative result as exhaustive. My searches returned hits that appear to reference this patent number but do not — these must not be conflated with it:

   • "2020 WL 5751232" — this is a Westlaw citation to Seagrape Investors LLC v. Tuzman, No. 19-cv-9736 (S.D.N.Y. Sept. 25, 2020), a securities-fraud case. It is not U.S. Patent 5,751,232, despite the coincidental string "5751232."
   • "the '575 patent" in several PTAB/district court documents refers to U.S. Patent No. 7,930,575 (Collabo Innovations v. Qualcomm) and, in another document, to U.S. Patent No. 5,748,575 (Intellectual Science & Technology v. Sony et al.). Neither is U.S. 5,751,232.
   • Walmart/tool catalog SKUs and an EP search-report family entry (US 5751232 A as a family member of EP 1330039) are likewise not litigation.

2. Search-tool coverage limits. I was not able to fully query PACER, Docket Navigator, Lex Machina, or the CAFC docket directly within the tool budget; the Unified Patents "caselist" endpoint returned only the patent record. A confirmed "no litigation exists" answer would require a run against those dedicated litigation databases (e.g., Docket Navigator, Lex Machina, Unified Patents' full litigation portal, CourtListener RECAP search on "5,751,232").

3. Practical context. The patent expired (anticipated 2014-07-29) and its predecessor/related family members (e.g., U.S. 5,748,575 in the Intellectual Science & Technology litigation) were asserted in the mid-2000s, but no suit naming 5,751,232 specifically surfaced. The fact that it is a Mitsubishi Electric encoding patent in the MPEG-adjacent video-compression space makes it plausible it was licensed rather than litigated, but I cannot confirm that from the available material.

Conclusion

Based strictly on the searches performed, I found no known litigation (no plaintiff, defendant, jurisdiction, case number, filing date, or outcome) involving U.S. Patent No. 5,751,232. Any assertion that such litigation exists would need to be verified against a dedicated litigation database (PACER, Docket Navigator, Lex Machina, or the Unified Patents litigation portal) before being relied upon. I am stating this as a negative finding based on limited coverage, not as an affirmative proof that no case was ever filed.

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

The first search returned confusable hits for a different patent ('575 patent = U.S. 7,930,575, Collabo v. Qualcomm). Let me confirm nothing exists for the '232 patent specifically.

Let me check for reexaminations or district court assertions, and confirm the fee-related expiration.

Proceedings overview

Total AIA trial proceedings on US 5,751,232: 0. The USPTO Open Data Portal returns no IPR, PGR, or CBM proceeding for this patent, and my independent web checks (Google Patents, the Unified Patents PTAB/patent portal, and general searches for the patent number and title) surfaced no petition, institution decision, FWD, or Federal Circuit appeal — there is likewise nothing to break down by status (0 active / 0 invalidated / 0 sustained / 0 settled / 0 institution-denied). The bottom line for a defendant is unusual and in your favor, but not because the patent is "hardened": this patent is expired (Google Patents lists a 2014-07-29 anticipated expiration and a legal status of "Expired - Fee Related"), assigned to Mitsubishi Electric Corp and inactive since well before the AIA trial regime matured. Any demand letter you receive citing this patent is a past-damages-only proposition at best, and the PTAB route is moot because the term has run.

No proceedings to report

I will not manufacture entries. To be explicit about what I checked and what I found: no proceeding number exists on file for this patent, so there is no petitioner, no panel, no institution decision, and no FWD to describe. Per the operating rule, the absence of PTAB activity is the finding.

---

⚠️ Naming trap — do not confuse this patent with the "'575 patent" line of cases

A search on "575" returns a large, active PTAB body of work that belongs to a different patent:

• IPR2019-01523 and IPR2019-01524 — Qualcomm Incorporated v. Collabo Innovations, Inc., challenging U.S. Patent No. 7,930,575 (a microcontroller standby/power-shutdown patent), not US 5,751,232. The Board's decisions in that family (e.g., the § 314(a) / multiple-petition analysis quoting the Trial Practice Guide) show the '575 patent was the subject of serial, duplicative Qualcomm petitions.
• Similarly, Mitsubishi Electric has appeared as patent owner in recent IPRs, e.g. a TCL petition against U.S. 9,635,656 (LTE scheduling-request claims) with a 2023-06-02 filing date and a February 2024 patent owner response — again, a completely different Mitsubishi patent.

If a demand letter or an IPR search report hands you "the '575 patent," verify the number digit-for-digit. US 5,751,232 has no IPR history; US 7,930,575 has a substantial one. Conflating them will produce a badly wrong defensive posture.

---

Strategic summary

Claim status: entirely untested at the PTAB, and now unenforceable going forward. Because no AIA trial was ever instituted, no claim of US 5,751,232 has been canceled and no claim has been sustained by the Board — the claims are "untested," not "hardened." The patent issued 1998-05-12 off an application filed 1996-08-28 (priority 1993-07-30) and, per the structured record, expired on 2014-07-29 with a legal status of "Expired - Fee Related." Practically, that means the patent owner cannot assert it against current or future products; only damages for past infringement that accrued before expiration, and even then only within the 35 U.S.C. § 286 six-year lookback from the date a suit is filed. You should confirm the expiration basis (natural term end vs. lapse for non-payment of maintenance fees) on USPTO PatentCenter/Global Dossier — Google's "Expired - Fee Related" label and "anticipated expiration" date are not always the same event, and it matters to the damages window.

Estoppel landscape: essentially empty, and largely academic. Since no petitioner ever obtained an institution decision, there is no § 315(e)(1) or § 315(e)(2) estoppel running against anyone on this patent, and no privies are estopped. Were the patent somehow still live, the available statutory vehicles would be narrow: it is a pre-AIA patent (application filed 1996), so PGR is unavailable (PGR applies only to applications filed on or after 2013-03-16), and CBM would fail because the claims are directed to DCT/variable-length video encoding (H03M/H04N), not a financial product or service. That leaves IPR — which requires an unexpired, assertable patent and, in a litigation posture, must be filed within one year of service of a complaint under § 315(b). Ex parte reexamination under 35 U.S.C. § 302 remains available at any time, including on an expired patent, but it is not an AIA trial and provides no estoppel leverage.

Pattern signals: silent. No petitioner has filed once against this patent, let alone twice; no patent owner appeal to the Federal Circuit exists; and there is no defensive aggregator (Unified Patents or similar) in the chain — Unified's public portal page for US-5751232-A carries only bibliographic and citation data, no IPR or litigation entry. The relevant nuance is that this patent sits in Mitsubishi Electric's 1990s video-coding family (the same specification line as U.S. 5,644,305 and cited as prior art in later Mitsubishi/ISO work such as the EP 1330039 and EP 1424856/EP 3515072 variable-length-coding families). The invention here — splitting orthogonal-transformed block data into 0-run-length data and coefficient data and run-length encoding each against the most probable value — was foundational to MPEG-era coding, which is exactly why one would expect IPR attention had the patent been asserted while live. It never attracted any.

---

Recommended next steps

1. Lead with expiration, not with validity. If you are a defendant, the dispositive point is that the asserted claims are expired (2014-07-29) and the patent carries an "Expired - Fee Related" status. The first move is a Rule 12(b)(6) / summary-judgment posture on non-infringement-on-its-face and on the § 286 damages window (six years back from filing), not an invalidity fight. Pull the maintenance-fee history and the reexamination certificate (if any) from USPTO PatentCenter and Global Dossier.
2. Verify ownership and any reexamination record. The structured record lists Mitsubishi Electric Corp as both original and current assignee. Confirm there is no unrecorded assignment to an NPE and no ex parte reexamination certificate that altered the claims — a reexam certificate would change the operative claim text you must chart against.
3. If (and only if) live assertion appears, the AIA-trial playbook is constrained: no PGR (pre-AIA), no CBM (non-financial subject matter), IPR only subject to the § 315(b) one-year clock. Cite the Board's own guidance on stacked/duplicative petitions — e.g., the Section 314(a) analysis in the IPR2019-01523 / IPR2019-01524 decisions (Qualcomm v. Collabo) — if a serial filer ever appears, but note that this reasoning comes from a different patent's record and is persuasive only.
4. Say plainly that there is no PTAB activity. The absence is itself the signal: a broadly asserted, economically meaningful patent of this vintage would attract IPRs. Here, none exist — consistent with an expired, dormant asset.

Sources consulted: US 5,751,232 on Google Patents (bibliographic data, expiration 2014-07-29, "Expired - Fee Related"); Unified Patents portal — US-5751232-A; USPTO PTAB E2E / PTAB Decisions portal; PTAB record for the different patent U.S. 7,930,575 (IPR2019-01523 / IPR2019-01524, Qualcomm v. Collabo Innovations) as retrieved via ptacts.uspto.gov petition documents. I found no FWD, no institution decision, and no CAFC opinion for US 5,751,232, and I did not locate any district court assertion of it — though my litigation search was search-engine-based and is not a substitute for a full PACER/Docket Navigator check if the stakes warrant it.

I'll research the assignment record for US 5751232 using available sources before drafting.

Assignment & Ownership Analysis — US 5,751,232

Sourcing caveat (read first): I was unable to retrieve a single recorded-assignment document for this patent. The USPTO Assignment Center (assignmentcenter.uspto.gov) is a client-side JavaScript application that did not return indexable records to the research tools available to me, so I cannot assert "the Assignment Center contains zero records" from first-hand inspection. What I can state is that no reel/frame, no conveyance type, and no post-issuance assignee beyond the original corporate owner surfaced in any indexed source — Google Patents legal events, Unified Patents' patent portal, Espacenet/EPO family data, Justia, or PatentLeaderboard. Per the operating rules, I do not fabricate reels/frames to fill the gap.

---

Inventors

All three are listed "all of Nagaokakyo, Japan" on the printed patent, which corresponds to Mitsubishi Electric's Nagaokakyo (Kyoto) development site — where the company's consumer VTR / digital video work was centered.

Unusual-pattern check — none found.

• Inoue is credited with 55 patents at Mitsubishi Electric (Patent Leaderboard), and continued co-filing digital-VTR patents with the same co-inventors well past this filing — e.g., US 5,744,306 (1998-04-14), US 5,907,660 (1999-05-25), US 6,141,485 (2000-10-31), US 6,337,948 (2002-01-08).
• The identical three-person team (Inoue / Onishi / Ishimoto) appears on sibling UK filings of the same digital-VTR program (GB 2,288,942; GB 2,295,946), and UK IPO journals list the same co-inventor set under "ISHIMOTO,JUNKO" on the earlier case and "SHINOHARA,JUNKO" on the later ones — consistent with a name change by the same inventor, though I flag this as an inference from co-inventor identity, not a confirmed legal fact.
• No inventor departed within 12 months of filing, and none shows a later assignment to a third party. The classic fire-sale precursor (inventor exodus → portfolio dump) is absent.

---

Original assignee

• Entity on the issued patent: Mitsubishi Denki Kabushiki Kaisha, Tokyo, Japan (7-3, Marunouchi 2-chome, Chiyoda-ku per the modern Mitsubishi Electric letterhead; the English rendering in current use is Mitsubishi Electric Corporation, TSE: 6503).
• Primary line of business: Diversified electrical/electronics manufacturer — energy systems, industrial automation, semiconductors/power modules, information systems, and consumer electronics, including the digital VTR / video recording line to which this patent family belongs. It remains a large operating company — not a holding or licensing vehicle.
• Product-embodiment question: The patent sits squarely in the digital-VTR high-efficiency (DCT + run-length / variable-length) codec family, a technology area Mitsubishi Electric commercialized in VTR-class products of that era. I cannot confirm from the record that any specific shipping product practiced the claims of this particular patent — that would require a claim chart against a product, which the assignment record does not provide.
• Current status: Operating, listed, no bankruptcy, no dissolution. The patent itself is "Expired – Fee Related" with an anticipated expiration of 2014-07-29 (Unified Patents records lifetime renewal fees of $0; Google Patents shows the 20-year term running to 2014-07-29). Either way, the asset is dead — it expired of term/lapsed and never entered an assertion chain.

---

Assignment timeline

Finding: no recorded post-issuance assignment was identifiable for US 5,751,232.

What the record shows:

• No reel/frame entries recovered. Neither Google Patents legal events nor any secondary aggregator (Unified Patents, Espacenet, Justia, PatentLeaderboard) shows a conveyance of this patent away from Mitsubishi Denki Kabushiki Kaisha.
• The only ownership link is the original one: inventors Inoue, Ishimoto and Onishi → Mitsubishi Denki Kabushiki Kaisha, executed on/around the filing of application 08/704,276 (filed 1996-08-28; Google Patents shows 1996-08-28, Unified Patents shows 1996-08-27, an artifact of time-zone normalization of the same event). This is the standard employee-inventor assignment and would ordinarily carry its own reel/frame at issue — I did not recover that reel/frame and will not invent one.
• Procedural context on the face of the document: the application is a divisional/continuation in a Mitsubishi digital-VTR family rooted in application 08/106,722 filed 1993-08-16 (which issued as US 5,499,144), with a claimed priority of 1993-07-30. Sibling US 5,739,968 is a divisional of the same parent. This is an internal corporate continuation practice, not a transfer.

If the Assignment Center in fact shows no records for this patent, that is the finding — it means the assignee of record has been Mitsubishi Denki Kabushiki Kaisha continuously from issue to expiry, and there was never an NPE in the chain.

---

Timeline diagram

timeline
    title Ownership of US 5751232
    1993 : Priority filing by Mitsubishi Denki
         : Parent app 08/106722 filed
    1996 : Divisional application filed
    1998 : Patent issued to Mitsubishi Denki
    2014 : Patent term expires
         : No post-issue transfer recorded

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No assignment from Mitsubishi Electric to any "IP / Holdings / Licensing / Ventures" entity appears anywhere in the record. The patent stayed with the operating manufacturer from 1998 to expiry (2014). There is no reel/frame to cite because no such conveyance exists in the indexed record.
2. Known asserter in the chain — not present. No assignee matches Acacia, Marathon, Intellectual Ventures, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, or any Unified Patents / RPX high-frequency plaintiff list. The only assignee of record is Mitsubishi Denki Kabushiki Kaisha.
3. Repeat correspondent across the chain — not present. With no assignment documents recovered, there is no correspondent of record to analyse. I have no attorney name, no firm, and no address to report — and I decline to guess at one. (Note the distinction from prosecution: the patent's prosecution counsel is a separate data point from assignment-recordation correspondents, and neither is available to me here.)
4. Cascading transfers — not present. Zero recorded transfers, so no chained-LLC sequence within 24 months (or any period).
5. Pre-litigation transfer — not present. No infringement suit naming this patent surfaced, and no assignment within 6 months of any such suit exists. The patent expired in 2014 with no assertion activity tied to it.
6. Bankruptcy fire-sale — not present. Mitsubishi Electric Corporation has not filed Chapter 7/11, and this patent was not part of any Kodak/Nortel/Polaroid-style asset sale.
7. Privateering — not present. No transfer to an NPE asserting on Mitsubishi's behalf; no SEC 8-K/10-K disclosure of a patent-transfer arrangement referenced this patent.
8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at expiry while still held by the original operating company.

---

Verdict

Insufficient data (original assignee still owns the patent; no post-issuance assignment identified).

Justification: No recorded assignment away from Mitsubishi Denki Kabushiki Kaisha could be identified for US 5,751,232 in Google Patents legal events, Unified Patents' portal, Espacenet family data, Justia, or PatentLeaderboard — the only ownership link evidenced is the original inventor-to-employer assignment tied to application 08/704,276 filed 1996-08-28, and the patent expired 2014-07-29 with no litigation or transfer. That absence of any reel/frame is itself the finding: all eight NPE signals read "not present," none is "unclear" on the merits (signal 3 is unrunnable for want of a correspondent, not adverse), and the operating assignee's continuous ownership plus the inventors' long post-filing tenure at the same company are the affirmative facts on the other side.

Verification links:

• USPTO Assignment Center (search patent 5,751,232): https://assignmentcenter.uspto.gov/ — and the legacy interface https://assignment.uspto.gov/patent/index.html
• Google Patents legal events: https://patents.google.com/patent/[US5751232A](/patent/US5751232A)/en#legal-events
• Third-party record confirming original/current assignee chain: https://portal.unifiedpatents.com/patents/patent/US-5751232-A

Recommendation for a follow-up analyst: because the Assignment Center could not be queried programmatically here, pull the record directly by patent number in the Assignment Center UI (and, if the original employee assignment matters, request the application's contents from the USPTO PatentCenter file wrapper for application 08/704,276) before treating "no records" as final. If the UI confirms zero hits, the file is closed as a captive operating-company portfolio asset.

I'll research the prior-art citations for US 5,751,232, starting with targeted searches.

Let me pull the front-page "References Cited" listing for this patent specifically.

I have the patent's front-page data. Now let me find the specific "References Cited" / "Patent Art" entries.

I'll research the prior art cited on this patent's face and any related art in the recursive/multi-stage run-length coding space before building the § 103 analysis.

Let me pull the specific prior-art references listed for this patent (Unified Patents lists 6 patent-art items and 2 NPL items).

Let me identify the patent family and the specific documents cited as prior art against this family.

Obviousness Analysis — U.S. Patent No. 5,751,232 (35 U.S.C. § 103)

0. Scope, sources, and confidence

What I could verify in this session: the patent's front page (title, inventors, assignee, appl. 08/704,276, filed 1996‑08‑28, issued 1998‑05‑12), the abstract, the admitted prior-art apparatus of FIG. 1 (shuffling → DCT → quantization → data division circuit 7 → VLC → buffer → rate control), and the JP family member JPH0818459A ("High efficiency encoder and high efficiency decoder"), whose published summary of the invention corroborates the multi-stage decomposition claim scope — see https://patents.google.com/patent/JPH0818459A/en.

What I could NOT retrieve (must be stated up front): the examiner‑cited art on the patent's face — i.e., the "(56) References Cited" list and the "Patent Art (6) / Non‑Patent Literature (2)" entries shown on https://portal.unifiedpatents.com/patents/patent/US-5751232-A. My tool budget expired before I could pull the patentimages PDF (https://patentimages.storage.googleapis.com/60/86/89/7e1805f8f8be80/[US5751232](/patent/US5751232).pdf).

Consequence: the grounds below are constructed § 103 theories that a challenger would plead, built from (a) the applicant's own admissions of prior art in the specification and (b) art demonstrably in the same field and timeframe. They are not represented as the art of record. Any item whose content I have not directly read is flagged.

Contradictions with the earlier-generated sections (flagged, not auto-corrected):

• Priority: Google Patents = 1993‑07‑30; Unified Patents = 1993‑07‑29. Also filing 1996‑08‑28 (Google) vs 1996‑08‑27 (Unified), and grant 1998‑05‑12 (Google) vs 1998‑05‑11 (Unified). These one-day offsets matter only at the margin of the § 102 boundary; they do not change the analysis below.
• The earlier section said the full claim set was not retrieved. That remains true. Claim 10 (verified) is the only independent claim text I can analyze with confidence; the decoder-side and dependent claims are addressed only at the level of the JP family counterpart's claim summary, which is expressly flagged as indirect.

---

1. The § 103 date problem (threshold issue)

This is a pre‑AIA case (priority 1993). The effective filing date — and thus the § 103 critical date — is the Japanese priority date of 1993‑07‑30 (or 1993‑07‑29) as to subject matter supported by the JP application, otherwise 1994‑07‑29 (parent Ser. No. 08/282,570).

Practical consequences:

• Art published before ~1993‑07‑30 is available under § 102(a)/(b) and § 103.
• MPEG‑1 (ISO/IEC 11172‑2:1993, published ~Aug. 1993) and JPEG (ISO/IEC 10918‑1:1994) are date-marginal for this patent. They were cited as "X" art against later Mitsubishi applications (see EP 1 424 856 A4 search report, http://data.epo.org/pise-server/rest/collections/lgpi/EP1424856A4.pdf), but against this patent a challenger would need their pre‑July‑1993 drafts (Committee Drafts circulated 1991‑92) to establish the 2‑D VLC/zig‑zag baseline as printed publication art. Drafts are usable but must be authenticated.
• If the JP priority were successfully attacked for lack of § 112 support for a given claim, the critical date slides to 1994‑07‑29, and MPEG‑1's published final text becomes cleanly available. That is a meaningful lever.

I flag this because it is the single biggest determinant of how strong any § 103 ground actually is.

---

2. Claim 10 — element breakdown and where the novelty actually sits

Claim 10 (as verified in the prior section) recites: (1) 1‑D scan → first 0‑run length data + first coefficient data; then a binary tree of six further dividers producing, at the leaves, third 0‑run length, fourth coefficient, second 1‑run length, fifth coefficient, fourth 0‑run length, sixth coefficient, first 2‑run length, and seventh coefficient data; (8) at least one VLC encoder for those eight streams; (9) a data synthesis circuit ordering the outputs.

Two observations drive the whole analysis:

(a) The first dividing stage is admitted prior art. The specification's FIG. 1 apparatus includes "a data division circuit 7 [which] separates the DCT data received from the quantizer 3 into 0‑run length data and coefficient data, and outputs them to a variable‑length encoder 4," followed by "2‑dimensional Huffman coding" using a coding table. That is an applicant admission that (i) DCT blocking, (ii) zig‑zag 1‑D scanning, (iii) separation of the scanned coefficients into zero-run-length data and non-zero coefficient data, and (iv) variable-length/Huffman encoding of those streams, were all known. Admissions in the specification are prior art for § 103 purposes and are often the most damaging evidence because they are unrebuttable.

(b) The only asserted advance is iteration. The JP family counterpart states the first invention as: 1‑D scan → separate into run-length data and coefficient data → "then the run-length data and coefficient data are further run-length encoded with the data having the highest occurrence probability as the run." The published "effects" section (JPH0818459A) frames the benefit purely as "further reduction of code amount in a reversible process." Invention 4 is expressly "run-length coding is further performed on the separated data a plurality of times," and inventions 5–9 add optimization details (reverse scan order, 1‑run first, different VLC tables, EOB′ codeword handling).

So the claim's delta over admitted art is: apply the same run-length decomposition recursively, keyed to the next-most-probable symbol (0, then 1, then 2), and re-multiplex the streams. That is exactly the fact pattern § 103 is designed to reach.

---

3. Candidate prior art (with qualification status)

Explicit warning: the last row is a trap. Both items appear in EPO search reports about the same family (EP 1 424 856, EP 3 515 072), but they are later art and cannot support a § 103 attack on this patent.

---

4. Grounds of rejection

Ground 1 — Admitted FIG. 1 apparatus in view of the two‑pass run‑length coding teaching

Combination: the admitted DCT/zig‑zag/0‑run‑coefficient‑separation/VLC apparatus, modified by the well‑known practice of run‑length coding the run‑length stream itself (and, symmetrically, the residual coefficient stream).

Why the combination is motivated: after the first pass, the residual distributions remain strongly peaked. Zero-run lengths cluster at 1‑2‑3; the coefficient stream, with zeros removed, is dominated by ±1; and so on. Applying the same RL/VLC machinery to those streams exploits the next probability peak and reduces average code length — the very result the spec claims ("further reduction of code amount"). A POSITA optimizing a VTR bitstream would (i) reuse the identical divider/encoder block already instantiated for pass one, (ii) reuse the existing VLC table infrastructure, and (iii) expect a purely lossless, reversible code‑amount reduction. That meets KSR's "known technique used to improve one device, and ... would improve similar devices in the same way," KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 417 (2007); see MPEP 2143, 2144.

Reasonable expectation of success: high. The operation is bit‑exact reversible; the only cost is an extra table and pipeline stage, both already present in the design.

Ground 2 — Admitted apparatus + the most‑probable‑symbol principle applied iteratively

Combination: the admitted apparatus + the conventional entropy‑coding principle that the symbol with the highest probability of occurrence should serve as the "run" symbol, applied successively to the residual symbol sets (0 → 1 → 2).

Motivation: the spec itself states the rationale ("run-length encoding ... by employing data of the highest probability of occurrence (for example 0)") — and that principle is the definition of efficient run-length coding. Once the principle is applied to the dominant symbol 0, a POSITA would mechanically reapply it to the now‑dominant symbol 1 in the residual, then 2. The JP counterpart confirms this is precisely the claimed scheme ("run-length coding by 1 is performed for the coefficient data" — invention 2). Repeating a known optimization step to the next‑most‑likely symbol is a design choice yielding a predictable result. MPEP 2144.04.

Ground 3 — Tree topology and stage count as design choice / obvious to try

The seven dividers of claim 10 define one particular binary tree over the alphabet. Nothing in the specification demonstrates that this specific tree is critical or produces an unexpected result; the spec frames the benefit generically as "further code amount reduction ... in a reversible process." Varying (a) how many stages, (b) which symbol is used as the run at each stage, and (c) the scan direction are expressly presented as alternative embodiments (JP inventions 4, 5, 6, 8). Where the patentee itself presents multiple interchangeable arrangements producing the same stated benefit, the selection of one arrangement is an obvious design choice.

Ground 4 — Dependent claims

The optimization features in the family record map to obviousness grounds:

• Different VLC tables for run-length data vs. coefficient data (JP invention 3) — routine table optimization; also directly a result of the two streams having different statistics.
• Reverse scan order for the second pass (JP invention 5) — matters of routine experimental optimization.
• 1‑run first, then 0‑run (JP inventions 6/7) — the same principle (Ground 2) applied to a different dominant symbol.
• EOB′ codeword distinct from EOB, plus a count of trailing 1s (JP invention 7) — analogous to escape/EOB signaling already used in 2‑D VLC of DCT coefficients (admitted art / MPEG‑1 style escape coding), applied by ordinary skill.

---

5. Motivation-to-combine statement (for the record)

A POSITA at the 1993 priority date, possessing the admitted DCT + zig‑zag + 0‑run/coefficient separation + VLC apparatus and the rate‑control feedback loop, would have been motivated to combine it with two‑pass run‑length coding because:

1. Articulated problem in the field: code‑amount reduction in a fixed‑rate digital VTR, with the spec's own rate‑control coefficient K explicitly tied to picture quality.
2. Known solution to a known problem: RL coding of a skewed integer stream was standard; the residual streams after pass one are themselves skewed integer streams.
3. Same elements, same way: reuse of the identical divider and VLC table hardware already disclosed in FIG. 1.
4. Predictable, verifiable result: lossless, bit‑exact, inverse‑process‑decodable.
5. No teaching away: nothing in the admitted art suggests that re‑coding the run‑length output would fail or would be inoperable; the only counter‑consideration (table overhead) is a routine trade‑off.

---

6. Where the patent could survive (rebuttal analysis)

A patentee would argue:

• Unexpected results (§ 716.02). The specification reports simulation results for luminance and chrominance, asserting "further code amount reduction ... compared with conventional high‑efficiency encoding apparatus." If the patentee can show the gain was unexpectedly large or non‑linear versus a naive two‑pass scheme, this is its best rebuttal. On the record I retrieved, the gains are asserted qualitatively; no comparative data table was recovered. This is the pivotal evidentiary battleground.
• Specific structural topology. The seven‑divider topology might be argued as a non‑obvious particular arrangement. Weak, given the spec treats the arrangement as one of several interchangeable embodiments — self‑created evidence of obviousness (Ground 3).
• Claim construction. The "circuits" are arguably means‑plus‑function under § 112 ¶ 6 (pre‑AIA). If construed narrowly to the disclosed algorithm and equivalents, the art must teach that algorithm — which Ground 1/2 do functionally.
• Priority/support. Not a rebuttal to obviousness, but a sword for the challenger (see § 1).

---

7. Recommended next steps to harden or break these grounds

1. Pull the actual "(56) References Cited" list from https://patentimages.storage.googleapis.com/60/86/89/7e1805f8f8be80/US5751232.pdf (cols. 1‑2). The examiner's own art is the strongest starting point, and I could not retrieve it within budget. The Unified Patents page lists 6 patent-art and 2 NPL items — identify each.
2. Obtain the JP family prosecution for JPH0818459A (and the corresponding JP application out of which US 08/282,570 arose) — JP examiners frequently cite older Japanese DCT/RL art for exactly this subject matter.
3. Authenticate pre‑July‑1993 MPEG‑1/JPEG drafts if the priority date survives challenge; otherwise use the Aug 1993 / 1994 publications if priority slips to 1994‑07‑29.
4. Verify the sibling patent. The earlier litigation section flagged US 5,748,575 (asserted in Intellectual Science & Technology v. Sony et al.) as possibly a related family member. If it shares the Inoue/Ishimoto/Onishi inventive entity and priority, its specification and prosecution history are directly relevant to both § 103 and § 112 analysis. I could not confirm it is a sibling — treat as a lead, not a fact.
5. Focus discovery on the "effects" evidence — the simulation data behind the code-amount claims — because that is the only non‑§ 103 escape route.

---

8. Conclusion

On the record retrieved, claim 10 of US 5,751,232 is vulnerable under § 103 as the repeated application of a known run‑length decomposition technique (the first stage of which the applicant admits was prior art in FIG. 1) to the residual run‑length and coefficient streams, keyed in turn to the most‑probable symbol (0, then 1, then 2). The motivation is supplied by the art‑recognized goal of reduced code amount in a fixed‑rate digital VTR, the result (lossless code-amount reduction) is predictable, and the specific seven‑divider tree is, by the patentee's own disclosure, one interchangeable arrangement among several — i.e., a design choice.

Caveats that materially qualify this conclusion: (i) I could not retrieve the examiner‑cited art, so these are constructed grounds, not the art of record; (ii) the priority date is the decisive variable — art from Aug. 1993 onward (notably MPEG‑1's final text) only becomes available if priority is defeated or the JP support is attacked; and (iii) the patentee's asserted but unevidenced simulation results are the strongest potential rebuttal and must be tested against the specification before any invalidity position is relied upon.

Sources: https://patents.google.com/patent/[US5751232A](/patent/US5751232A)/en; https://portal.unifiedpatents.com/patents/patent/US-5751232-A; https://patents.google.com/patent/JPH0818459A/en; http://data.epo.org/pise-server/rest/collections/lgpi/EP1424856A4.pdf; http://data.epo.org/pise-server/rest/collections/lgpi/EP3515072A1.pdf; https://patentimages.storage.googleapis.com/60/86/89/7e1805f8f8be80/US5751232.pdf (not retrieved within budget).