US 10212422

I'll search for the specific patent number and any related docket activity, then compile the summary from the authoritative patent text provided.

The searches returned useful bibliographic and litigation data. The provided patent text does not include the claims section (it ends mid-description at Table 14), so I'll search for the actual claim language before summarizing the independent claims.

US Patent 10,212,422 — Summary

Note on sourcing: The full patent description was provided in the request. The claims section was not included in the provided text, so the independent-claim language below is sourced from RPX Insight's record of the patent (verified against the Google Patents record). Bibliographic and status data come from the Google Patents record provided and live search results.

Bibliographic data

• Title: Method and apparatus for encoding video, and decoding method and apparatus
• Patent number: US 10,212,422 B2 (application no. 15/801,423; pre-grant publication US 2018/0070084 A1, published 2018-03-08)
• Inventors: Sung Chang Lim, Hui Yong Kim, Se Yoon Jeong, Suk Hee Cho, Jong Ho Kim, Ha Hyun Lee, Jin Ho Lee, Jin Soo Choi, Jin Woong Kim, Chie Teuk Ahn (10 named inventors)
• Original assignee: Electronics and Telecommunications Research Institute (ETRI)
• Current assignee (per Google Patents): Ideahub Inc. (reassigned from ETRI via assignments recorded in 2023)
• Filing date: 2017-11-02 (continuation-style filing)
• Priority date: 2010-08-17
• Issue date (grant): 2019-02-19
• Status: Active; anticipated expiration 2031-08-17 (with patent-term considerations)
• Classifications: H04N 19/124 (quantisation), H04N 19/176 (block-level adaptive coding), H04N 19/119 (adaptive subdivision), H04N 19/70 (syntax), among others

Abstract

The invention relates to a video encoding method and apparatus for setting and encoding quantization parameters, and to a video decoding method and apparatus for decoding and setting quantization parameters, in a video codec that uses blocks having various sizes and depths as encoding/decoding units. Quantization parameters are adaptively set according to block information (e.g., size/depth of the coding, prediction, or transform unit) to improve quantization/inverse quantization, deblocking filtering, entropy coding, inter prediction, rate control, and rate-distortion optimization.

Independent claims (plain-language overview)

The patent has 7 claims; the independent claims are 1, 4, and 7 (decoding method, encoding method, and computer-readable medium).

• Claim 1 – Decoding method: A decoder reads, from the bitstream, (a) "base block size information" defining a base block size and (b) "residual presence information" telling whether the current decoding unit contains a residual signal. Only if a residual is present does the decoder parse a residual quantization parameter for that unit; it then derives the unit's quantization parameter from that residual value and performs inverse quantization. Key limitation: if the current decoding unit is smaller than the signaled base block, its quantization parameter is forced to be the same as that of all other decoding units within the same base block (i.e., one QP per base block, inherited by all sub-blocks).

• Claim 4 – Encoding method: Mirror-image encoding method: the encoder determines and signals the base block size; determines and signals residual presence for each encoding unit; determines a quantization parameter; quantizes using it; and, when a residual is present, computes and encodes a residual quantization parameter. Same limitation as claim 1: units smaller than the base block all share the base block's quantization parameter rather than receiving their own signaled QP.

• Claim 7 – Non-transitory computer-readable medium: A storage medium carrying a bitstream generated by the encoding method of claim 4 (same steps and same base-block QP-sharing limitation).

Dependent claims 2, 3, 5, and 6 further limit the methods (exact text not captured in the sources retrieved; I do not have authoritative wording for them and did not fabricate it).

Litigation / docket search results (with caveats)

• Google Patents flags the patent family as having litigation (Darts-IP family link), and the '422 patent has reportedly been designated as essential to the VVC/H.266 standard in the Access Advance VVC patent pool.
• CAFC: The Federal Circuit oral-argument archive lists appeal 2024-1684, IdeaHub Inc. v. Unified Patents, LLC. I could not confirm from the retrieved sources that this specific appeal concerns the '422 patent (Ideahub owns many related codec patents), so do not treat that as established.
• Reexamination: A Unified Patents article (URL dated 2026-08-31) states that on August 28, 2026, Unified Patents filed an ex parte reexamination against US 10,212,422. ⚠️ This reported date is after the stated current date (April 26, 2026) in this session; per the operating rules I am reporting the live search result as-is, but I flag this inconsistency rather than reconciling it.
• I found no specific CAFC 2026 docket tied by number to the '422 patent in the searches performed.

Uncertainty statement

• The claim text above comes from RPX Insight's database, not directly from the USPTO Patent Center in this session; it is consistent with the specification but I could not verify the dependent claims' exact wording.
• The CAFC appeal 2024-1684's subject patent(s) could not be verified as including the '422 patent.
• The August 2026 reexamination filing date conflicts with the session's stated current date and should be treated with caution.

Based on my searches, here is what I can report regarding litigation involving US Patent 10,212,422 (the '422 patent). I want to be precise about what is verified versus what is not.

Known proceeding (USPTO, not a court case)

Ex parte reexamination – Unified Patents, Inc. v. Ideahub Inc.

• Proceeding type: Ex parte reexamination filed at the USPTO (this is an administrative proceeding, not district-court or CAFC litigation)
• Requestor: Unified Patents, LLC (described as representing itself; in-house counsel TJ Murphy and Jessica L.A. Marks are identified as counsel)
• Patent owner: Ideahub Inc. (current assignee of the '422 patent)
• Filing date per source: August 28, 2026
• Status: Filed; no outcome identified in the sources retrieved
• Source: Unified Patents article, "IdeaHub H.266 / VVC codec patent challenged" (https://www.unifiedpatents.com/insights/2026/8/31/another-ideahub-video-codec-patent-challenged)
• ⚠️ Date inconsistency flag: That Unified Patents article is dated August 31, 2026, which is after today's stated date of April 26, 2026. Per the operating rules I report the live search result as-is, but I flag that the reexamination filing date (August 28, 2026) post-dates the current date in this session. Treat the filing date with caution.
• The article also notes the '422 patent "has been designated as essential to the VVC standard in the Access Advance's VVC patent pool" and that a related patent was asserted in Europe (no European case number or court identified in the source).

Family-level litigation flag (no case specifics retrieved)

Darts-IP family record:

• Google Patents displays a "Family has litigation" flag for US10212422B2, linking to a Darts-IP family record (family ID 45839092, https://patents.darts-ip.com/?family=45839092...).
• I could not retrieve from that link any specific U.S. case number, parties, court, or filing date for litigation naming the '422 patent. So while the family is flagged as having litigation, I cannot verify that any U.S. district court case asserts the '422 patent itself.

Unverified CAFC lead (do not treat as established)

IdeaHub Inc. v. Unified Patents, LLC – Federal Circuit appeal 2024-1684:

• The prior analysis noted this appeal appears in the Federal Circuit oral-argument archive. However, I could not confirm from any retrieved source that appeal 2024-1684 concerns the '422 patent specifically. Ideahub owns numerous related codec patents (e.g., US 10,904,539, which Unified also challenged in a separate January 31, 2025 ex parte reexamination), so this appeal may concern a different patent. I am listing it only as an unverified lead, not as confirmed litigation involving the '422 patent.

District court litigation

My searches did not return any district court complaint, case number, plaintiff/defendant pair, or docket entry specifically identifying US 10,212,422 as an asserted patent. I therefore have no verified U.S. district court litigation to report for this patent as of the current date.

Bottom line

Uncertainty statement: I could not verify any district court litigation for the exact patent number US 10,212,422; the only directly relevant proceeding identified is the Unified Patents ex parte reexamination, whose reported filing date conflicts with the session's stated current date. The CAFC appeal 2024-1684's involvement of this patent is unconfirmed. For authoritative, up-to-the-minute case data, I would recommend checking Unified Patents' Portal (portal.unifiedpatents.com), USPTO Patent Center (for the reexamination control number), and PACER.

Proceedings overview

Total AIA trial proceedings (IPR/PGR/CBM) on US 10,212,422: zero. The canonical USPTO Open Data Portal block in this prompt shows no AIA trial proceedings on file for this patent, and targeted web searches (PTAB dockets, Unified Patents, CAFC, district-court databases) surfaced no IPR, PGR, or CBM petition against the '422 patent. The only post-grant challenge on record is an ex parte reexamination filed by Unified Patents on 2026-08-28 — three days before today's date — which is a 35 U.S.C. § 302 proceeding, not an AIA trial. Bottom line for a defendant: the patent is entirely untested before the PTAB in AIA trials — no claim has been canceled or sustained, no IPR estoppel attaches to anyone, and the claims remain fully in force while the just-filed reexamination plays out.

No AIA trial proceedings to report

• Type: None — no IPR, PGR, or CBM petition has been filed against US 10,212,422.
• Status: The structured USPTO ODP data ("PTAB proceedings on file") is empty for this patent; web searches returned no petition numbers, no institution decisions, and no Final Written Decisions referencing the '422 patent.
• Judge panel / grounds / institution / FWD / settlement / appeal: N/A — nothing to report. I did not and will not invent proceeding numbers.

Adjacent but distinct activity — ex parte reexamination (not an AIA trial):

• On 2026-08-28, Unified Patents filed an ex parte reexamination request against US 10,212,422 (Unified Patents insight article, dated 2026-08-31, linked below). The patent is designated as essential to the VVC/H.266 standard in the Access Advance pool; a related patent is asserted in Europe. Unified is represented by in-house counsel TJ Murphy and Jessica L.A. Marks. As of today (2026-08-31), the USPTO has not yet made a substantial-new-question (SNQ) determination, and the reexamination control number (90/xxx) had not appeared in the retrieved sources — I could not confirm it and did not guess. This proceeding is outside the AIA-trial scope of this task but is the single most important post-grant development for a defendant.

Correction to prior analysis — CAFC appeal 2024-1684 does not involve the '422 patent. The Federal Circuit's Rule 36 judgment in IdeaHub Inc. v. Unified Patents, LLC, No. 2024-1684 (affirmed 2026-04-10), states on its face that it is an appeal from PGR2022-00044 — a proceeding against US 11,122,274 (a directional intra-prediction patent), not US 10,212,422. That appeal should be excluded from any defense analysis of the '422 patent.

• Defensive value: Because no IPR has been instituted or decided, there is no § 315(e) estoppel and no PTAB claim-construction or validity finding binding on any future challenger. Every § 102/§ 103 ground remains available to a defendant. Conversely, the patent owner cannot point to any PTAB victory — this patent is not "hardened"; it is simply unchallenged at the Board.

Strategic summary

Claims: CANCELED vs. SUSTAINED vs. UNTESTED. All seven claims of US 10,212,422 (independent claims 1, 4, and 7, with dependents 2–3 and 5–6) are UNTESTED in any AIA trial proceeding. No PTAB panel has ever construed a claim of this patent, considered prior art against it, or issued a Final Written Decision. The ex parte reexamination filed 2026-08-28 could narrow or eliminate claims, but as of today no SNQ order has issued and no Office action on the merits is public.

Estoppel landscape. There is no IPR, so 35 U.S.C. § 315(e)(2) estoppel binds no one. Any defendant facing assertion of the '422 patent today can raise any § 102 or § 103 combination it can support — including art and arguments that overlap the pending reexamination (note: ex parte reexamination is not an estoppel-generating AIA trial, and its examiner-driven record does not collaterally estop later litigants, though a defendant should be alert to whether Unified Patents' members are the real parties in interest for any future IPR). Because the priority date is 2010-08-17, the full pre-AIA § 102(a)/(b)/(e)/(g) and § 103 art universe is available.

Pattern signals. There is no repeat-petitioner pattern on this patent — no petitioner at all, yet. But the surrounding signal is loud: Unified Patents has an active VVC/H.266 challenge program and, within three days of today, added the '422 patent to it via reexamination (a forum it has increasingly favored as IPR institution rates have fallen under the USPTO's 2025–2026 discretionary-denial reforms). Unified previously prevailed against IdeaHub in PGR2022-00044 (US 11,122,274 canceled; affirmed under Rule 36 at the CAFC on 2026-04-10), showing that IdeaHub's VVC-essential portfolio is being systematically probed. If Unified later converts its reexamination effort into an IPR — or if an operating-company member steps forward — this patent is a plausible target.

Recommended next steps

1. Treat the ex parte reexamination as the live threat, and monitor it weekly. Source: Unified Patents announcement (https://www.unifiedpatents.com/insights/2026/8/31/another-ideahub-video-codec-patent-challenged). Watch for (a) the USPTO's SNQ determination and reexamination control number (90/xxx, not yet public in retrieved sources), (b) the first Office action, and (c) any IdeaHub response. If the examiner adopts Unified's grounds, expect claims to be rejected in the first action; the patent owner then has the burden to overcome them or amend.
2. Since no IPR estoppel exists, preserve full freedom to run prior art. If you are a defendant in (or expect) district-court litigation, do not assume the reexamination will dispose of the case — reexamination and litigation run in parallel. Marshal § 102/§ 103 art against the actual claim language of claims 1, 4, and 7 (parse the residual-QP / base-block-sharing limitation carefully — that is the claim's heart) and be ready to file an IPR within one year of service if litigation materializes; there is no earlier-filed IPR to trigger § 315(b) complications.
3. If you are IdeaHub's licensee or a VVC implementer facing a license demand, the calculus differs: no claim has been weakened by the PTAB, so the patent is legally presumed valid and enforceable today; the reexamination is early-stage and provides no immediate negotiation leverage beyond the uncertainty it creates.
4. Confirm the absence of PTAB activity directly once you have credentials: check PTAB E2E (https://ptab.uspto.gov) and the USPTO ODP API for "10212422" — the structured data in this prompt shows no AIA trials, and my searches corroborated that, but E2E is the authoritative real-time docket if the reexamination later spawns an IPR.
5. Track the reexamination's progress against the timeline: ex parte reexaminations typically see an SNQ order within ~2–3 months of filing (expect a decision roughly Q4 2026), followed by a first Office action; unlike IPR, there is no statutory one-year trial clock, so do not expect fast resolution.

I'll research the assignment record for US 10,212,422 from the USPTO Assignment Center and cross-reference sources.

Ownership & Assignment Record — US Patent 10,212,422

⚠️ Retrieval limitation (read first). I was not able to open the USPTO Assignment Center or the legacy Assignment Search UI directly in this session; my step budget was exhausted on cross-reference searches. The reel/frame numbers below are those exposed in the Google Patents legal-events record for US10212422B2 (which mirrors Assignment Center data), and I have marked clearly which items I could not confirm. I did not retrieve any assignment correspondent (recording attorney/firm) for this patent and will not invent one. All items I could not verify are labeled as such.

Inventors

All ten named inventors, with employer at time of filing as determinable from the EPO/Turkish register entry for the EP family member (EP 3402194, same title, same inventors, same KR 10-2010-0079104 priority of 2010-08-17) — where correspondence addresses are recorded as "C/O ETRI", Daejeon, Korea:

Pattern notes.

• All rights were assigned to ETRI; no inventor appears as an assignee at any point.
• No inventor-departure signal. All ten inventors were still the named inventors of record at the 2017-11-02 continuation filing and at the 2019-02-19 grant, with ETRI holding title. I found no evidence of a mass inventor exodus near filing.
• The real pattern is the continuation strategy, not inventor mobility. The '422 is one link in a chain ETRI built from the single 2010-08-17 disclosure (PCT → national phase → US 2013/0136172 → US 9,838,691 → US 10,212,422 → US 10,839,106 / 10,939,106 / 11,265,546 / 11,601,649 / 12,088,807). Splitting one 2010 specification into at least eight granted U.S. patents creates multiple independently assertable and SEP-declarable assets — the structure a monetization buyer wants.
• ⚠️ Unverified: whether any inventor personally benefited from the 2023 Ideahub transfer. Korean employee-invention law (and the ETRI/IP pool structure) typically routes compensation through the institute. Not established here.

Original assignee

Electronics and Telecommunications Research Institute (ETRI), Daejeon, Republic of Korea.

• Entity type / products: Government-funded, non-profit ICT research institute founded 1976; it does not ship consumer products embodying the claims. Its output is R&D, standards contributions and patent licensing (HEVC/VVC standardization history; licensor in HEVC Advance, Sisvel's VP9/VVC platform, Access Advance's VVC pool, and Via LA's ATSC 3.0 pool per retrieved pool documents).
• Current status: Operating — no bankruptcy, dissolution or acquisition event found. ETRI remains an active licensor and standards participant as of the retrieved 2024–2026 sources. (It remains a licensor alongside Ideahub in the Sisvel VP9 programme, i.e. ETRI retained other patents even after selling this family.)
• The '422 issued to ETRI on 2019-02-19 (assignee of record at issue). Title later left ETRI (see timeline).

Assignment timeline

Four recorded events appear in the Google Patents legal-events record for the '422. Two reel/frame citations are recoverable from the face of the corrective assignments themselves; the others were not retrievable this session.

1. 2017-11-02 (executed date not retrieved) / recorded 2017-11-02 — Reel/Frame not retrieved

• Conveyance: Assignment ("Assignment of Assignors' Interest")
• Assignor: all ten inventors — KIM, Hui Yong; LEE, Jin Ho; AHN, Chie Teuk; CHO, Suk Hee; CHOI, Jin Soo; JEONG, Se Yoon; KIM, Jin Woong; KIM, Jong Ho; LEE, Ha Hyun; LIM, Sung Chang
• Assignee: Electronics and Telecommunications Research Institute (ETRI)
• Correspondent: not retrieved — ⚠️ cannot assess recurrence.
• Context: Internal/corporate capture of employee invention rights by the employer — recorded contemporaneously with the 2017-11-02 continuation filing (ETRI's standard practice of re-recording the inventor assignment for each continuation in the chain).
• ⚠️ Note: the recordation date is 2017; the execution date of the underlying inventor assignment is not exposed in the source I used. ETRI practice and the family structure suggest execution circa 2010–2012, but I cannot confirm it.

2. 2023-04-14 / recorded 2023-04-14 — Reel 063321 / 0082

• Conveyance: Assignment (subsequently corrected)
• Assignor: recorded as IDEAHUB — this entry is erroneous; the conveying party name was later corrected to ETRI (see event 3).
• Assignee: IDEAHUB (later corrected to IDEAHUB INC.)
• Correspondent: not retrieved — ⚠️ cannot assess recurrence.
• Context: Transfer-to-asserter / monetization purchase — ETRI's title to the '422 passes to a Korean patent licensing company. Execution date not retrieved; only the 2023 recordation is documented.
• ℹ️ Timing context: ETRI's transfer of this video-codec family is recorded in 2023, whereas Ideahub's acquisition of ETRI's separate MPEG-DASH/streaming family was pleaded as occurring June–August 2018 (Helios Streaming/Ideahub complaint, ¶29, D. Del.). The two appear to be separate transactions.

3. 2023-04-21 / recorded 2023-04-21 — Reel 063917 / 0456

• Conveyance: Correction ("Corrective Assignment to correct the conveying party name previously recorded at Reel 063321 Frame 0082")
• Assignor: Electronics and Telecommunications Research Institute (ETRI) (confirming the assignment)
• Assignee: IDEAHUB (itself corrected later — see event 4)
• Correspondent: not retrieved — ⚠️ cannot assess recurrence.
• Context: Documentation cleanup of a bulk transfer — the conveying party on the original record was mis-identified as the buyer rather than the seller, a paperwork defect typical of high-volume portfolio conveyances.

4. 2023-07-26 / recorded 2023-07-26 — Reel/Frame not retrieved (the corrective references Reel 063917/0456)

• Conveyance: Correction ("Corrective Assignment to correct the receiving party data name from IDEAHUB to IDEAHUB INC., previously recorded on Reel 063917 Frame 0456")
• Assignor: Electronics and Telecommunications Research Institute (ETRI)
• Assignee: IDEAHUB INC. (final named owner)
• Correspondent: not retrieved — ⚠️ cannot assess recurrence.
• Context: Change of name / correction of assignee legal name only — no change in beneficial ownership.

Post-2023-07-26: No further U.S. assignment is recorded in the sources retrieved. The chain currently terminates at IDEAHUB INC. (consistent with Google Patents' "Current Assignee: Ideahub Inc."). I found no record assigning the '422 to Helios Streaming, LLC or to any pool vehicle — note that Helios holds an exclusive licence (not title) in the sibling streaming family, which is a different structure from this family's.

Timeline diagram

timeline
    title Ownership of US 10212422
    2010 : Priority filing by ETRI inventors
    2017 : Continuation filed by ETRI
         : Inventor rights assigned to ETRI
    2019 : Patent issued to ETRI
    2023 : ETRI assigns to Ideahub
         : Two corrective assignments filed
    2026 : Unified files ex parte reexam

NPE / troll-pattern signals

Verdict

NPE — high confidence.

Title to US 10,212,422 left an operating, product-less research institute and now sits — via reel 063321/0082 (recorded 2023-04-14), corrected at reel 063917/0456 and again on 2023-07-26 — with Ideahub Inc., an entity expressly identified as "an NPE" by Unified Patents, described by RPX as a Korean patent-monetization firm, parent of litigating NPE Helios Streaming, and management-affiliated with Glocom/SIPCO. That is two strong signals: a transfer from an operating assignee to a licensing-only entity (Signal 1), and a known asserter as current owner (Signal 2). The absence of anonymized-shell trappings and of any verified U.S. suit on this specific patent does not change the classification — the '422 is a VVC-pool-declared asset held purely for licensing revenue, and it was selected by Unified as worth a 2026 reexamination precisely because of that posture.

Verify independently at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (enter 10212422 or application 15/801,423); legacy mirror at https://assignment.uspto.gov/patent/index.html → Patent Search → "10212422". Pull the PDFs behind Reel 063321 Frame 0082 and Reel 063917 Frame 0456 to obtain the execution dates and the recording correspondent — those two data points are the only material gaps in this analysis, and I did not guess at either.

I'll research the cited references for US Patent 10,212,422. Let me start by searching for the patent's citation data from multiple sources.

The initial searches returned the patent's own text and forward citations, but not the backward citations (references cited by the examiner). Let me search more specifically for the references cited on the face of the patent.

The Google Patents snippets aren't surfacing the backward citation list. Let me try Espacenet and other patent databases which typically expose the examiner-cited references.

I've reached the search-step limit, so I'll now consolidate what the searches actually confirmed versus what I could not retrieve, and give you the most defensible prior-art analysis I can — with explicit flags on anything not verified.

Direct answer to the search task

What I searched: USPTO-focused queries for the backward-citation list ("References Cited") of US 10,212,422, plus Google Patents, Espacenet, and family-tracing queries keyed to the literal number 10212422 (and its publication number US10212422B2, application 15/801,423).

What I could not retrieve: The examiner-cited references on the face of the patent (the "References Cited" / "Patent Citations" section). The Google Patents page for US10212422B2 was returned but the search snippets exposed only the specification text and forward citations ("Cited By"), not the backward citation table. I did not obtain the USPTO Patent Center or Espacenet citation record, and I will not invent citation numbers. Any specific reference list below that is not tied to a retrieved source is explicitly labeled as analysis, not as a confirmed examiner citation.

What the searches did confirm (verified facts)

• The patent: US10212422B2, "Method and apparatus for encoding video, and decoding method and apparatus," Electronics and Telecommunications Research Institute (ETRI); granted 2019-02-19 from application 15/801,423 filed 2017-11-02; priority date 2010-08-17. (Google Patents record, provided text.)
• Same-family earlier publications (all verified in search results):
  • US20130136172A1 — published 2013-05-30, same title and same ETRI inventors/priority (verified in the citation table of the US20110279644A1 page).
  • US9838691B2 — granted 2017-12-05, same family (verified).
  • US20170223355A1 — published 2017-08-03, same family (verified).
  • WO2012023806A3 — PCT publication, 2012-05-10, listed under "Also Published As" for this family (verified via the JP6615923 page).
  • Other family members: US10939106B2, US10827174B2, US11265546B2, US11601649B2, US12088807B2, US20190141325A1, US20220150499A1, KR101650464B1, KR20150023576A, JP6615923B2 and related JP patents, EP4391532A2 (verified via the JP6615923 "Also Published As" table).
• Family chain implied by the data: the '422 patent is a continuation in a chain anchored to a 2010-08-17 priority application (PCT published as WO2012023806 → US national phase US20130136172A1 → continuations US9838691B2 → US10212422B2).

Prior-art analysis for claims 1, 4, and 7

Recall the independent claims (per the earlier sections of this analysis): claim 1 (decoding method), claim 4 (encoding method), claim 7 (computer-readable medium/bitstream). The claim heart is: signaling a base block size; signaling residual presence information; parsing/encoding a residual quantization parameter only when a residual is present; and forcing every decoding unit smaller than the base block to inherit the base block's single QP rather than receiving its own signaled QP.

1. Same-family earlier publications — the closest "prior art" candidates, but with a critical §102 caveat

• US20130136172A1 (published 2013-05-30) and US9838691B2 (granted 2017-12-05) disclose essentially the same specification as the '422 patent (the family's QP-setting/QP-prediction disclosure). If the '422 patent validly claims priority to the 2010-08-17 application (35 U.S.C. §§ 119/120), these are not prior art under §102(a)(1)/(a)(2) — they are the same inventive entity's own earlier-filed disclosures with a proper priority chain.
• The §102 exposure arises only if the '422 claims contain new matter not supported by the 2010 application. Any claim limitation whose written-description support first appears in the 2017 continuation filing would have an effective filing date of 2017-11-02, making US20130136172A1, US9838691B2, US20170223355A1, WO2012023806A3, and the Korean publication (KR20120016980A-type family members) genuine §102(a)(1)/(a)(2) prior art against that new matter. This is precisely the kind of attack a §302 reexamination requestor (e.g., Unified Patents, per the earlier litigation summary) would mount. This is the single most important prior-art angle for this patent, and I can state it with confidence from the verified family data — but the actual rejection grounds in the pending reexamination are not public in any source I retrieved.

2. Foundational standards art (analytical, not a confirmed citation)

• ITU-T H.264 / ISO/IEC 14496-10 (MPEG-4 AVC), published well before 2010: the macroblock-layer mb_qp_delta and slice-level slice_qp_delta syntax, plus per-macroblock QP inheritance/change rules, is the canonical disclosure of (i) signaling QP changes only when needed and (ii) block-level QP deltas relative to a slice/base QP. It is the most natural §102/§103 anchor against the residual-QP-plus-base-QP concept of claims 1, 4, and 7 — though it does not disclose the HEVC-style quadtree CU/TU hierarchy with a signaled base block size that the '422 claims center on. It would more likely support a §103 obviousness combination than a clean §102 anticipation.
• JCT-VC / HEVC early drafts and test-model documents (2010–2011): the April 2010 JCT-VC Test Model under Consideration (JCTVC-A205) and the first HEVC working drafts describe CU/TU partitioning with QP handling at coding-tree level. Their dates straddle the 2010-08-17 priority date, so their viability as §102 art depends entirely on the priority-date validity analysis above (they are only usable for claims whose effective date is 2011 or later).

3. Claim-by-claim anticipation mapping (analytical framework — no confirmed citations)

Uncertainty statement

• I could not retrieve the actual examiner-cited references for US10212422B2 in this session; the specific citation list on the face of the patent is therefore not reported rather than guessed. The H.264 and JCT-VC items above are analytical candidates, not confirmed citations from the patent's file wrapper.
• The pending ex parte reexamination (per the earlier litigation summary, filed by Unified Patents, reportedly 2026-08-28) would, once its control number and first Office action become public, reveal the actual art the requestor is pressing — that is the highest-value next retrieval, along with the Google Patents "Citations" tab and USPTO Patent Center for the 15/801,423 file wrapper.
• To complete this task authoritatively, I recommend: (1) pull the PDF front page of US10212422B2 from Patent Center (the "References Cited" box), (2) open the Google Patents US10212422B2/en page's Citations tab, and (3) check Espacenet's "Cited documents" for the publication number. I did not reach those pages' citation tables within the allowed steps.

Obviousness Analysis — US Patent 10,212,422 (35 U.S.C. § 103)

0. Sourcing caveat (read first)

The user's instruction asks me to rely on "the results from the Prior Art section of this page." No Prior Art section was included in the materials provided to me — the supplied text is the patent specification (which ends mid-Table-14), the bibliographic/summary sections, and the litigation/PTAB sections. I therefore reconstructed the prior-art set from three sources and flag each item's provenance:

1. Search-verified references (surfaced in this session): JCTVC-E198 (Kobayashi, "Sub-LCU Level Delta QP Signaling"), JCTVC-F046 (Chono, "Efficient Binary Representation of cu-qp-delta Syntax for CABAC"), JCTVC-H1003 (HEVC WD6), and JCTVC-D258 (cu_qp_delta addition to the coding-unit layer) — these appeared as non-patent citations in the related Mediatek patent US 9,912,950 and in a PTAB exhibit, not in the '422's own record.
2. Well-known standards from training knowledge (H.264/AVC, MPEG-2, H.263, HEVC WD1/JCTVC-A003, Samsung's JCTVC-A124 proposal) — flagged as training-derived; I could not re-verify their full text this session.
3. The '422 specification itself (its Tables 1–14 disclose the syntax the inventors considered novel: qp_hierarchy_depth, log2_qp_unit_size, minCUDQPsize, lcu_qp_level, cu_qp_delta, transform_unit_qp_delta, residualDataPresentFlag).

I did not retrieve the examiner's "References Cited" list for the '422, and I did not retrieve the Unified Patents reexamination request (the task's stated current date of April 26, 2026 predates the reexamination's reported August 28, 2026 filing date — an inconsistency I flag rather than reconcile). Where a reference's content is asserted from training knowledge rather than this session's search results, I say so explicitly.

1. Claims at issue

Claim text is from RPX Insight (verified against the specification; the Google Patents text provided ends before the claims). Independent claims 1 (decoding method), 4 (encoding method), and 7 (bitstream medium) share one inventive core:

(a) decode/encode base block size information defining a base block size;
(b) decode/encode residual presence information for a decoding/encoding object unit (does it have a residual?);
(c) only if a residual is present, decode/encode a residual quantization parameter for that unit;
(d) derive the unit's QP from that residual value and (inverse-)quantize;
(e) if the object unit is smaller than the base block, its QP is forced equal to the QPs of all other units inside the same base block (one QP per base block, inherited by every sub-block).

Dependent claims 2, 3, 5, and 6 were not retrieved; I do not analyze them by number, only the class of dependent limitations disclosed in the specification (QP prediction from left/above/prior blocks, median prediction, SPS/PPS/slice-header QP, collocated-block prediction, Z-scan ordering, prediction-mode-similarity, etc.).

Effective prior-art date. The patent claims priority to 2010-08-17 (KR filing). For § 103 purposes the primary art universe is what a PHOSITA knew before that date. Several highly relevant JCT-VC documents (E198, F046, H1003, D258) postdate the priority date; they are usable as § 103 art only if the priority claim fails (e.g., new matter added in the 2017 continuation) — I flag them as secondary/fallback references below.

2. Prior art references

3. Legal framework (Graham / KSR)

Obviousness under § 103 asks: (1) the scope and content of the prior art; (2) differences between the prior art and the claims; (3) the level of ordinary skill; and (4) secondary considerations — evaluated with KSR's flexible, common-sense motivation inquiry (no rigid TSM requirement; "obvious to try," design incentives, and known problem-solution pairing can suffice). A PHOSITA here is a video-codec engineer familiar with block-based hybrid codecs (H.264/AVC) and the HEVC standardization effort circa 2010, competent in syntax design, entropy coding, and rate control.

4. Combination 1 (primary) — H.264/AVC (R1) + HEVC WD1/Samsung hierarchical block structure (R3/R4) → claims 1, 4, 7

Claim-element mapping:

Why a PHOSITA would combine: The JCT-VC effort (April 2010 onward) was explicitly chartered to build on H.264/AVC's tools while replacing the fixed 16×16 macroblock with a variable-size hierarchical block tree. The single most obvious adaptation when carrying H.264's per-block QP-delta tool into a hierarchical codec is to (i) keep one QP per node at some chosen granularity ("base block" — the LCU or a signaled sub-LCU size), (ii) have all descendant blocks inherit it (a consequence of the tree, not an invention), and (iii) continue to signal only a delta relative to a predictor. Each of the three independently claimable features — block-size signaling in high-level syntax, residual-gated QP delta, and QP inheritance by sub-blocks — is either expressly taught by R1 or structurally inherent in R3/R4. The combination is a textbook "known technique (QP delta signaling) applied to a new context (variable-size block tree)" with a predictable result (per-region quantization control at lower signaling cost). Under KSR, this is the classic obvious substitution/adaptation.

Claim 4 (encoding) is the mirror image and maps identically; claim 7 (bitstream medium) adds nothing patentably distinct because a bitstream generated by an obvious method is itself obvious (the medium limitation is conventional).

Weakness to anticipate: no single pre-priority reference discloses the exact triad (signaled base-block size + residual-gating + forced inheritance). The proponent must articulate the combination explicitly — which the standardization record supports, since JCT-VC documents from Jan–Nov 2011 (R5–R8) show multiple independent teams converging on precisely this design within a year. That convergence is itself evidence the design was obvious and not a pioneering leap.

5. Combination 2 — H.264/AVC + residual-presence gating (coded block flag) → element (b)/(c)

Element (c) — "decode the residual QP only when residual presence info says a residual exists" — is the most attackable limitation. H.264 already conditions mb_qp_delta on the macroblock being coded (skipped MBs carry no QP delta), and coded_block_pattern/coded_block_flag were the standard, long-known mechanisms for signaling residual presence (R1; also MPEG-4 Part 2 cbpy/coded flags). A PHOSITA designing a hierarchical codec would immediately see that sending a QP delta for a block with no residual wastes bits and produces no effect on reconstruction — the motivation to gate QP-delta signaling on the coded-block flag is self-evident bit-overhead reduction, with a reasonable expectation of success because the two mechanisms were already co-present in H.264 macroblock syntax. This element alone does not support patentability.

6. Combination 3 — base-block-size signaling (minCUDQPsize-type syntax) → element (a)/(f)

Element (a)/(f) — signaling a "base block size" and forcing smaller units to share its QP — corresponds directly to H.264's fixed 16×16 macroblock granularity (one QP per MB; sub-partitions share) and to the HEVC LCU concept in R3/R4 (the LCU is the natural "base block"; SPS already signals its size). Choosing the QP granularity (LCU vs. sub-LCU vs. a signaled minimum QP-unit size) is a routine rate-control design parameter. MPEG-2 (R2) and H.263 (R9) likewise applied one quantizer step per macroblock to all contained sub-blocks. Setting the granularity by a syntax element in PPS/slice header — exactly what H.264 does for slice QP and what the '422 spec itself places in SPS/PPS/slice header — is the obvious way to make a known design choice (QP granularity) adaptive. Nothing in element (a)/(f) is a new kind of signaling; it is a new value of a known parameter.

7. Combination 4 — dependent-feature limitations (predictors)

The specification's dependent limitations are all known H.264/JCT-VC prediction mechanisms, making them obvious over R1 alone or R1+R3:

• QP prediction from a previously decoded/encoded block (left/above/Z-scan): H.264 predicts a macroblock's QP from the previous macroblock's QP; JCT-VC proposals (R5–R8) extended prediction to left/above CUs. The "Z-scan order" is the standard traversal of a quad-tree (already in R3/R4's coding tree).
• Median of adjacent QPs: median prediction was already the H.264 motion-vector prediction technique; applying the same median-of-neighbors idea to QPs is an obvious analog (the '422 spec itself offers the median as an alternative to identification-flag signaling, i.e., a designer's obvious fallback to avoid signaling overhead).
• QP from SPS/PPS/slice header: squarely R1 (pic_init_qp_minus26, slice_qp_delta).
• Prediction from a collocated block in a previously decoded picture / reference-picture block indicated by motion info: collocated-block derivation is standard in H.264 temporal prediction; reusing it as a QP predictor is an obvious extension.
• Prediction by prediction-mode/intra-direction similarity: selecting predictors by mode similarity is a known refinement pattern in H.264 intra coding.

Caveat: the specific left/above CU-QP prediction and cu_qp_delta machinery appear in post-priority documents (R5–R8). If the 2010 priority date holds, these can only support obviousness of the combination logic, not as standalone § 102/§ 103 references; if the priority chain fails (new matter in the 2017 continuation), R5–R8 become directly available and the obviousness case is substantially stronger.

8. Motivation, expectation of success, and secondary considerations

• Standardization context supplies motivation as a matter of record: JCT-VC's express purpose was to improve coding efficiency over H.264 while reusing its proven tools; multiple teams filed near-identical QP-per-CU proposals within months (R5–R8), showing a recognized problem (QP signaling overhead with larger blocks / finer granularity) and a narrow design space of solutions.
• KSR "obvious to try" / design need: as block sizes grew from 16×16 to 64×64 LCUs, fixed per-picture QP was too coarse and per-mini-block QP too costly; the obvious middle ground — one QP per base block with residual-gated deltas — is exactly the claimed solution, arrived at independently by several groups.
• Reasonable expectation of success: all components (QP delta, QP prediction, coded-block-flag gating, block-size SPS syntax, tree-based inheritance) were proven in H.264/MPEG-2; integrating them into a new tree structure is engineering, not invention.
• Secondary considerations: I found no evidence of long-felt need, unexpected results, or copying. The patent's designation as "VVC-essential" cuts against non-obviousness here: the claimed technique became part of a mainstream standard, consistent with it being the ordinary design answer rather than a surprising one. (Note: simultaneous independent development by multiple JCT-VC contributors is affirmative evidence of obviousness.)
• Validity presumption: claims 1, 4, 7 are untested in any AIA trial (per the PTAB section above); a challenger must still overcome the statutory presumption with clear and convincing evidence, but the record here — H.264's per-block QP-delta + the HEVC tree structure — is a strong combination for that burden.

9. Bottom line

Strongest § 103 combination: ITU-T H.264/AVC (R1) — disclosing per-block residual-gated QP delta signaling, QP prediction from prior blocks, slice/PPS QP syntax, and uniform QP across a macroblock's sub-blocks — in view of HEVC WD1 (R3) and/or the Samsung proposal (R4), disclosing the hierarchical LCU/CU/TU tree, SPS block-size signaling, and tree-inherent QP inheritance. That combination maps every element of independent claims 1, 4, and 7 and renders the claimed subject matter obvious to a PHOSITA under Graham/KSR. If the priority date is successfully attacked (continuation/new-matter), the case becomes materially stronger with R5–R8 (JCTVC-E198/F046/D258/H1003) as direct art disclosing cu_qp_delta in HEVC.

Most vulnerable limitation is the residual-presence gating of QP-delta parsing (element (c)) — a self-evident bit-saving optimization over H.264's coded-block-pattern/skip machinery. Least vulnerable (and the point a patent owner would stress) is the exact combination of signaled base-block size + forced QP equality among all sub-units — no single pre-priority reference states that precise rule verbatim, so the analysis necessarily rests on a two-reference combination and the standardization record's demonstration that the design was the obvious, convergent solution.

Uncertainty statement: The examiner's cited references for the '422, the dependent claims' text, and the reexamination request's art were not retrieved this session; R3/R4 content is asserted from training knowledge; and the reported reexamination filing date (2026-08-28) postdates this task's stated current date (2026-04-26). Any formal opinion should verify R1/R3/R4 texts against JCT-VC/ITU archives and pull the '422 file history (USPTO Patent Center) before relying on this analysis.