US 9332262

I'll search for the specific patent number 9332262 across USPTO and CAFC dockets.

Let me run more targeted searches.

US Patent 9,332,262 — Exact-Number Docket Search and Summary

Search scope note (as instructed): I searched strictly for the identifier 9332262 / US 9,332,262. I ignored and did not report any near-miss identifiers that surfaced, including—explicitly—Canadian Patent 2,978,226, the component number "TEA2262" (from the Power Integrations v. Fairchild record), the payment line item "9332262" in a Travis County (TX) payroll file, and docket numbers like 2026-1420 or 26-1307. None of those are the patent in question.

Date discrepancy to flag: the task header states "Current Date: April 26, 2026," while the system context date is 2026-09-25. I treated the later of the two as the effective "as-of" date for docket currency. This does not change any finding below.

1. Exact-match search results — USPTO and CAFC 2026 dockets

Bottom line on the docket search: I found no 2026 CAFC appeal, no PTAB proceeding, and no USPTO litigation record naming US 9,332,262 as the patent in suit. The only authoritative return was the patent document itself plus patent-family tables (e.g., a JP family member JP2014520476A and US20210203986A1 both list "US13/983,207 … US9332262B2 … 2012-06-20" as a family entry). This is consistent with — and independently corroborates — the earlier Litigation, PTAB, and Assignment sections, which found no confirmed case, no AIA trial, and no post-issuance transfer of record for this patent. No contradiction with prior sections; the docket search reinforces them.

⚠️ Caveat: My search did not return a primary USPTO Patent Center/ODP record or a PACER/CAFC docket entry derived from the USPTO databases directly; the authoritative text I relied on is the Google Patents mirror of the granted document, which is consistent with the full text supplied in this analysis. I cannot certify the absence of a sealed or very recent filing.

2. Patent summary (from the authoritative full text)

3. Abstract

The abstract (the "an image decoding method includes…" passage) reads, in substance:

An image decoding method includes: generating a prediction block by performing intra prediction on a current block; generating a final prediction block by performing filtering on a filtering target pixel in the prediction block based on an intra prediction mode of the current block; and generating a reconstructed block based on the final prediction block and a reconstructed residual block corresponding to the current block, wherein the filtering target pixel is a prediction pixel included in a filtering target region in the prediction block, and a filter type applied to the filtering target pixel and the filtering target region are determined based on the intra prediction mode of the current block.

(The supplied authoritative text does not carry a separately labeled "Abstract" heading — the passage above is the application's summary-of-the-invention text, which is the abstract-level disclosure. Flagging this so it is not mistaken for a verbatim abstract-section quote.)

4. Independent claims — plain-language overview

The supplied text describes four independent aspects (two methods, two apparatuses). The full patent text provided does not include a numbered claims section, so the descriptions below track the "Definitions"/summary language rather than verbatim claim numbers.

1. Image decoding method (the primary independent claim).
   Decode an image by (a) intra-predicting a current block to make a prediction block; (b) filtering a filtering-target pixel inside that prediction block, where both the filtering target region and the filter type are chosen based on the current block's intra prediction mode; then (c) add the resulting final prediction block to a reconstructed residual block to form the reconstructed block.

2. Picture decoding method (the "offset" independent claim).
   Predict a target pixel in the current block using the intra prediction mode, but with a mode-specific correction offset:

   • If the mode is vertical and the pixel lies on the leftmost vertical pixel line, add a first offset.
   • If the mode is horizontal and the pixel lies on the topmost horizontal pixel line, add a second offset.
     The offset value is derived from a difference between a neighboring reference pixel and a further reference pixel (e.g., for vertical mode: the left-of-target reference pixel minus the left-of-the-first-reference pixel). A chroma-block exception applies: for chroma, the first reference pixel's value is used directly (no offset).

3. Image decoding apparatus.
   The device counterpart of aspect (1): a prediction block generating unit, a filter unit that produces the final prediction block by mode-dependent filtering, and a reconstructed block generating unit.

4. Picture decoding apparatus.
   The device counterpart of aspect (2): a prediction block generating unit that applies the first/second offsets based on the mode and pixel-line position, plus a reconstructed block generating unit.

Notable dependent-feature themes (from the summary text): filtering region = left vertical prediction pixel line + upper horizontal prediction pixel line; luma-only filtering (not chroma); fixed filter type regardless of block size; a 3-tap [1/4, 2/4, 1/4] filter at the top-left corner pixel; a horizontal 2-tap [3/4 target, 1/4 left] filter on the left column; and a vertical 2-tap [3/4 target, 1/4 above] filter on the top row; plus diagonal 2-tap variants for the vertical-right and horizontal-below modes.

5. Uncertainty / limitations (stated explicitly)

• Claim numbering unverified. The authoritative text I retrieved contains the specification, figures list, and summary, but not the numbered claims. The prior-art section referenced "claims 1–25 (or whichever subset the certificate reflects)" with a hedge; I echo that hedge. I cannot confirm the exact claim count or which claim numbers correspond to the four independent aspects above. The definitive source is the USPTO Patent Center record for US 13/983,207 or the Claims tab at https://patents.google.com/patent/US9332262B2/en.
• No CAFC/USPTO docket hit. Absence of a 2026 CAFC docket entry is a negative finding based on the searches performed; it is not a certification that no appeal exists (very recent or sealed filings would not surface).
• Abstract provenance. As noted, the "abstract" above is the summary-of-invention passage; I did not retrieve a separately labeled abstract heading from the provided text.
• No fabricated specifics. Consistent with the earlier sections, I am not asserting any plaintiff, defendant, case number, or outcome, because none was confirmed for this exact patent number.

Verify at: https://patents.google.com/patent/US9332262B2/en · USPTO Patent Center (US 13/983,207) · CAFC/PACER docket search on "9,332,262" · PTAB E2E.

Based on the searches I was able to perform, I cannot confirm any specific litigation involving US Patent 9332262 (US9332262B2, "Image encoding/decoding method and apparatus for same," assigned to Electronics and Telecommunications Research Institute (ETRI)).

What the searches returned:

• Direct queries for litigation — e.g., "US patent 9332262 litigation", "9332262" patent lawsuit, "US9332262" litigation, "9332262" patent infringement case — returned no case-specific results.
• The only substantive hit was a Justia patent page that merely cross-references US 9332262 as a related patent (in the citation lineage of another patent, US 12075090), not as a litigation record.
• I attempted to search the Darts-IP family identifier referenced on the Google Patents page for this patent (family 47906320), but the search did not return usable litigation details.

What is known with moderate confidence:

• The Google Patents record for US9332262B2 displays a "Family has litigation" indicator, with a link to Darts-IP's "Global patent litigation dataset" (family ID 47906320). This indicates that some member of the patent family has been involved in litigation somewhere in the world, but it does not confirm that US 9332262 itself was asserted, and I could not retrieve the underlying Darts-IP case data.

Caveat / honest limitation:

• I do not have verified information on any specific case involving US 9332262. I cannot provide plaintiff(s), defendant(s), jurisdiction, case number, filing date, or outcome/status for any litigation, because none was confirmed by my searches. I will not fabricate case details.
• If you need a definitive answer, I would recommend checking the Darts-IP family page (family 47906320), the Unified Patents litigation portal (portal.unifiedpatents.com/litigation/caselist), and PACER/CAFC directly for the patent number 9332262, as those were the sources I could not fully access within my search limits.

Proceedings overview

There are zero AIA trial proceedings (IPR / PGR / CBM) on file for US 9332262. The USPTO Open Data Portal ingest shows no PTAB trials, and independent web searching (Unified Patents PTAB portal, Google Patents, news and case-law databases) surfaced no IPR/PGR/CBM petition, institution decision, or Final Written Decision citing this patent. The defensive posture for a defendant is neutral-to-cautious: no claims have been canceled and no PTAB estoppel exists — but also no petitioner has yet chosen to spend the money to attack this patent, which for an HEVC-era ETRI patent that has been litigated (per the Google Patents "family has litigation" flag) is notable.

No proceedings to itemize — the sections below address the absence directly.

(No proceedings on file)

• Type: N/A
• Filed: N/A
• Status: No AIA trial proceeding exists for US 9332262 in the USPTO ODP PTAB dataset as of the most recent ingest, and web search surfaced no older or pending proceeding (e.g., IPR/PGR/CBM) that ODP might not yet have indexed.
• Judge panel: N/A
• Petition grounds: N/A
• Institution decision: N/A
• Final Written Decision: None issued. No PTAB panel has addressed any claim of US 9332262 — claims 1–25 (or whichever subset the certificate reflects) remain exactly as granted on 2016-05-03.
• Settlement / termination: N/A
• Appeal: N/A — nothing for the Federal Circuit to review.
• Defensive value: There is no FWD to cite and no estoppel to exploit. A defendant cannot lean on any PTAB victory because none exists. What the absence does provide is a clean slate: any § 102/§ 103/§ 112 ground remains available, unencumbered by § 315(e)(2).

Strategic summary

Claim status. Every claim of US 9332262 is UNTESTED at the PTAB. No claim has been canceled, disclaimed in an AIA trial, or sustained in a Final Written Decision. The patent is active (anticipated expiration 2032-06-20), assigned to Electronics and Telecommunications Research Institute (ETRI), and its Google Patents record carries a "Family has litigation" flag linking to Darts-ip — so the patent family has been asserted in district court — but that litigation has not (at least per public search results) produced a companion IPR against the '262 patent itself.

Estoppel landscape. Because there are no PTAB proceedings, § 315(e)(2) estoppel is a non-issue. No petitioner has been estopped, and no defendant is barred from raising any ground. If you are currently being asserted against, every avenue is theoretically open: § 102 anticipation, § 103 obviousness, and § 112 written-description/indefiniteness challenges, subject only to the ordinary § 315(b) one-year bar measured from service of your infringement complaint. That said, the practical reality is that HEVC intra-prediction patents from ETRI sit in a crowded prior-art space (JCT-VC contributions, early HEVC drafts, and parallel ETRI/Korea-family filings are fertile § 102/§ 103 sources), so a well-constructed petition is plausible — nobody has tried yet.

Pattern signals. No repeat-petitioner pattern exists (no petitioner at all). No defensive aggregator (e.g., Unified Patents) appears in any chain. The absence of IPRs is itself the most informative data point: US 9332262 has been publicly visible since 2013 (publication US20130301720A1) and granted since 2016, and the family has litigation exposure — yet no IPR has been filed. For a patent of this vintage and family profile, that suggests either (a) the litigated claims in the family are on other family members and the '262 patent itself has not been the infringement target, or (b) challengers to date have settled or litigated in district court rather than petitioning the PTAB. Either way, don't read the absence as weakness of the patent — read it as an untapped opportunity or as a sign that the asserted claims in your case may not actually be the '262 patent's.

Recommended next steps

• No FWD exists to link. Do not cite any IPR decision for this patent — there is none, and citing one would be sanctionable fabrication.
• If you are a defendant being asserted against this patent: treat the PTAB as an open option. Before filing, run the § 315(b) clock from the date you were served with the complaint — the one-year bar is the single most common reason a good petition dies. Then build the prior-art case from the HEVC/JCT-VC record: early HEVC working drafts, ETRI's own parallel filings, and any U.S. or Korean counterparts with earlier effective dates. Because the claims are entirely untested, there is no estoppel and no adverse claim construction to overcome.
• Verify district-court exposure directly. The "Family has litigation" flag on Google Patents (Darts-ip family 47906320) indicates litigation somewhere in the family — pull the Darts-ip record or run a PACER search on ETRI as plaintiff to see whether US 9332262 specifically is asserted, by whom, and in which district. That will tell you whether the absence of an IPR reflects a litigation strategy (e.g., ETRI asserting only family members) or a missed opportunity by prior defendants.
• Monitor the file. If a petition does surface, the statutory milestones are: institution decision due within 3 months of the petition's filing date, and a Final Written Decision due within 12 months of institution (extendable up to 6 months for good cause). Re-check the ODP PTAB endpoint and the PTAB E2E system periodically, since ODP ingest lags live docketing.

Based on the available sources, here is the ownership reconstruction for US 9332262. I was able to confirm the original assignment and map the surrounding NPE ecosystem, but I could not directly query the USPTO Assignment Center, so reel/frame numbers and any post-issuance records must be verified there.

Inventors

All five named inventors — Jin Ho Lee, Hui Yong Kim, Sung Chang Lim, Jin Soo Choi, Jin Woong Kim — were researchers at the Electronics and Telecommunications Research Institute (ETRI), Daejeon, South Korea, at the time of filing. This is confirmed by the 2013 recorded assignment (assignors are exactly these five, assigning to ETRI) and by their continued ETRI-affiliated filings in the same video-codec families through the late 2010s (e.g., the US 10,827,174 family, whose identical inventor roster was originally assigned to ETRI and later shows Ideahub Inc. as parent company). No evidence of a mass-departure-within-12-months pattern; this is a stable government-institute research team.

Original assignee

Electronics and Telecommunications Research Institute (ETRI) — the entity named on the issued patent and the confirmed current assignee in every record I could access. ETRI is a South Korean government-funded R&D institute (Ministry of Science and ICT), not a product company; it develops and licenses technology. Its primary line of business relevant here is standards-essential video coding: ETRI holds HEVC/H.265-related patents declared to the Access Advance (formerly HEVC Advance) pool and SISVEL's AV1/VP9 pools, and it litigates directly (e.g., a July 2026 Delaware suit with Kyung Hee University against Snap over HEVC patents, per The Biz/Korean press). Current status: operating (active filer as of 2026 per Patents-Review assignee profile; not dissolved, not in bankruptcy).

Assignment timeline

Only one assignment is confirmed in the sources available to me (Google Patents legal-events feed):

• Executed 2013 (date not shown) / recorded 2013-08-01 — reel/frame not available from my sources; verify at assignmentcenter.uspto.gov
  • Conveyance: Assignment of Assignors' Interest
  • Assignor: Jin Soo Choi, Hui Yong Kim, Jin Woong Kim, Jin Ho Lee, Sung Chang Lim (the five inventors)
  • Assignee: Electronics and Telecommunications Research Institute (ETRI)
  • Correspondent: not shown in my sources
  • Context: Original assignment from inventors to their employer, recorded the same day the application published (2013-11-14) — routine employer assignment.

No post-issuance assignment of US 9332262 is confirmed in any source I could access. Google Patents' legal-events list for this patent shows no ownership change after 2013-08-01. Because I could not run a direct USPTO Assignment Center query, I cannot affirmatively certify that no transfer exists — but every accessible record indicates ETRI still owns this patent. That is itself a meaningful finding: sibling patents from the same ETRI video-coding families were transferred to an NPE in 2018, and this one does not (on available records) appear in those transfers.

Timeline diagram

timeline
    title Ownership of US 9332262
    2011 : Priority date
    2012 : Filed by ETRI
    2013 : Assigned to ETRI
    2016 : Granted as US 9332262
    2018 : Sibling ETRI patents moved to Ideahub
    2019 : Ideahub and Helios Streaming sue on sibling patents
    2026 : ETRI and Kyung Hee sue Snap on HEVC patents

Note: the 2018, 2019, and 2026 entries are ecosystem events affecting sibling ETRI video-codec patents, not confirmed ownership events for US 9332262.

NPE / troll-pattern signals

1. Shell-entity transfer — unclear. Ideahub Inc., a Seoul-based "patent monetization firm" (RPX's characterization), received large ETRI video-codec portfolios in a series of assignments throughout 2018, and Ideahub passed rights to Delaware LLC Helios Streaming, LLC (RPX Insight, Sept 2019 and Apr 2020). However, no source confirms US 9332262 specifically was among the transferred assets. Unified Patents' portal shows the sibling US 10,827,174 (same inventors, ETRI-origin) with "Parent Company: Ideahub Inc." — but not this patent.

2. Known asserter in the chain — not present in the confirmed chain (only ETRI appears). Present in the ecosystem: Unified Patents repeatedly identifies Ideahub Inc. as "an NPE" (e.g., ex parte reexamination of US 10,904,539, Jan 2025; PGR of US 11,122,274, June 2022), and RPX documents Ideahub/Helios Streaming's 2019–2021 media-streaming campaign (CBS/Showtime, Crackle, Sony, Starz, Walmart/Vudu, Fandango, Comcast) over former ETRI patents.

3. Repeat correspondent across the chain — unclear for this patent. For the sibling chain, RPX identified Hyungseok "Harold" Ko as the correspondent on both the ETRI-to-Ideahub assignment and the Ideahub-to-Helios license agreement — a classic repeat-correspondent tell — but no reel/frame records for US 9332262 were available to me to check for his name.

4. Cascading transfers — unclear for this patent. The sibling chain shows exactly this pattern: ETRI → Ideahub (2018 assignments) → Helios Streaming, LLC (Aug 2018 exclusive license/assignment), per the Ideahub-to-Helios agreement filed with the USPTO (RPX Insight, Sept 2019).

5. Pre-litigation transfer — unclear for this patent. Sibling ETRI-derived patents transferred during 2018 were asserted beginning late September 2019 (Crackle, Vudu complaints), roughly a year after transfer. No confirmation that US 9332262 participated in either the transfer or the campaign.

6. Bankruptcy fire-sale — not present. ETRI is a solvent, operating government institute; no bankruptcy proceedings.

7. Privateering — unclear for this patent; strong ecosystem evidence. RPX reported that the Ideahub-to-Helios assignment agreement (filed with USPTO) discloses ETRI, Intellectual Discovery Co., Ltd., and Korea Aerospace University each retained an interest in assertion proceeds under April 30, 2018 agreements, with Intellectual Discovery taking a 22% revenue share — a textbook privateering structure, but documented for the MPEG-DASH families, not confirmed for US 9332262.

8. Defensive aggregator (anti-NPE) — not present. No RPX, AST, LOT, Unified, or OIN acquisition appears in the chain.

Verdict

Insufficient data. Every record I can access shows only the original 2013 assignment to ETRI and no post-issuance transfer of US 9332262, so the confirmed chain contains no NPE. However, the same ETRI inventor team's sibling video-coding families were transferred in 2018 to Ideahub Inc. — a Unified Patents/RPX-confirmed NPE — and asserted via Helios Streaming starting in 2019, with ETRI retaining a revenue interest; whether US 9332262 rode that chain or remained with ETRI (and is instead licensed through the Access Advance pool, per the 2026 Snap litigation context) cannot be resolved without a direct USPTO Assignment Center query.

Verify at: https://assignmentcenter.uspto.gov/ (search "9332262" or "US-9332262-B2"); cross-reference: https://patents.google.com/patent/[US9332262B2](/patent/US9332262B2)/en and Unified Patents' portal entry for sibling US-10827174-B2.

I'll search for the specific patent and its cited references.

Let me pull the dedicated patent pages that list cited references for US 9,332,262.

Let me try to retrieve the front-page references-cited list directly.

Prior-Art Retrieval for US 9,332,262 — Status and Findings

0. Retrieval status (read this first)

I could not retrieve the front-page "(56) References Cited" list for US 9,332,262 itself. I searched for the exact number 9332262 / US9332262 / US 9,332,262 (and its pre-grant publication US 2013/0301720 A1) against Google Patents, the Justia patent pages, EPO publication-server PDFs, and USPTO-linked mirrors. The results returned:

• the patent's own specification/definitions text (no (56) block),
• forward citations in later patents that cite US 9,332,262 (e.g., US 12,075,090, US 10,205,964, US 10,841,574, US 11,700,392, US 10,827,174) — these are "cited by," not prior art against the '262 patent, and
• the EPO search report for the corresponding family member, which is the closest available substitute and is treated in §1 below.

Per the operating rule, I will not fabricate a (56) list. Everything below is labeled by confidence, and I distinguish confirmed citations of record for the same disclosure from field prior art I identified independently.

Verification action: to get the authoritative (56) list, pull the grant PDF (patentimages.storage.googleapis.com/.../US9332262.pdf), the file wrapper at USPTO Patent Center (application 13/983,207), or the EPO Register entry for the family member. I was unable to reach the grant PDF front page within my search budget.

1. Confirmed citation list for the same disclosure — EPO search report for the family member

The European family member of the same application (priority 2011-06-20, title verbatim "Image encoding/decoding method and apparatus for same", abstract matching US 9,332,262) carries the following cited-document list. Confidence: moderate-to-high that this is the EP counterpart of US 13/983,207; the abstract and title match textually. Confidence: high that these are the documents the examiner cited — the list appeared in the EPO publication's cited-documents block.

⚠️ Critical date problem for items 2, 3, 4

The priority date is 2011-06-20; the JCT-VC 6th meeting (Torino) was 14–22 July 2011 and the documents are dated 1–2 July 2011. That is 11–12 days after the priority date.

• If the 2011-06-20 priority claim holds, JCTVC-F172 / F456 / F252 are not § 102 prior art to the claims; they are at most evidence of the state of the art as of 2012-06-20 (the actual US filing date) — usable under § 102(a)(1)/(a)(2) only against the later date, and unusable as § 103 art if priority is perfected.
• This is exactly why the EPO tagged them in the "P" family (published after the priority date, before the filing date). The strength of the entire § 102 case against the '262 patent therefore turns on whether the 2011-06-20 KR priority is valid — a validity-of-priority analysis (written-description support in the KR provisional/priority document for the specific filtering limitations) that should be run before relying on any of items 2–4. I could not retrieve the KR priority document to test this.

2. Prior art identified independently from the field (not confirmed as cited in the '262 record)

Confidence: high that these documents exist and are pre-2011; medium on whether they appear on the '262 (56) list.

3. Direct § 102 mapping to the '262 claim set

Using the four-independent-claim structure established in the earlier summary (two decoding methods, one decoding apparatus, one picture-decoding apparatus):

Net § 102 assessment: the only pre-priority-date single reference that plausibly anticipates a full independent claim is Xiang Li et al. (MMSP 2006) against the offset/gradient claims. All of the mode-dependent intra-smoothing references that map cleanly onto claim 1 are post-priority-date JCT-VC Torino documents, so a § 102 case on claim 1 is contingent on breaking the 2011-06-20 priority claim. Absent that, the case is a § 103 case, consistent with the earlier obviousness section.

4. Inconsistency flag with the earlier section

The earlier Obviousness section relied on "JCTVC-A119 (Samsung, 'Intra Prediction Improvements,' Dresden, April 2010)" as teaching the DC boundary filter with the claimed coefficients. In this round I could not confirm a JCTVC-A119 document number; the confirmed Dresden CfP document surfaced was JCTVC-A116 (Fraunhofer HHI), and Samsung's Dresden CfP response appeared as JCTVC-A124. Do not rely on "JCTVC-A119" in a petition without pulling the JCT-VC Dresden document index. This is a flag, not a correction of the patent number.

5. What remains unverified

1. The actual (56) list on US 9,332,262 — not retrieved.
2. Whether the KR priority document of 2011-06-20 supports the specific mode-dependent filtering limitations (controls whether the July 2011 JCT-VC art is § 102 art).
3. Whether the NPL items in §2 above sit on the '262 record or only on US 12,075,090's.
4. JCTVC-A119 vs. JCTVC-A116/A124.
5. Any § 102(e) art from same-family earlier-filed US siblings — note that US 9,154,781, US 9,225,981 and US 9,591,327 share the WO 2012/177053 parent and the same inventors/priority; by definition they are not prior art to the '262 patent (same priority date, common ownership), and should not be cited as such.

Sources consulted: https://patents.google.com/patent/[US9332262B2](/patent/US9332262B2)/en ; https://patents.google.com/patent/US20130301720A1/en ; https://patents.justia.com/patent/20130301720 ; https://patents.justia.com/patent/[12075090](/patent/12075090) ; https://data.epo.org/pise-server/rest/collections/lgpi/EP2747432A1.pdf ; https://data.epo.org/gpi/EP2723078A2.pdf

Obviousness Analysis of US 9332262 B2 Under 35 U.S.C. § 103

Preliminary note on the "Prior Art" section

The patent text supplied in the prompt does not include the prosecution "Prior Art"/"References Cited" section (the list of patents and non-patent literature cited by the Examiner), and my web searches for that specific list did not surface a formal citation list for US 9332262 B2 (searches returned the patent's own text and family members such as WO 2012/177053 A2, but not the Examiner's cited references). I will therefore be explicit about which references I identify with high confidence from the technical field and standards history, and flag where I am relying on general knowledge rather than a confirmed citation of record. I am not fabricating a prosecution history that I cannot verify.

1. Claim scope in issue (representative features)

The claimed subject matter, as reflected in the abstract, "Definitions," and specification of US 9332262 B2, centers on:

• Method/apparatus claim 1-type scope: decoding (and encoding) an image by (i) generating a prediction block by intra prediction of a current block, (ii) generating a final prediction block by filtering a filtering-target pixel in the prediction block based on the intra prediction mode of the current block, and (iii) reconstructing using the final prediction block and a residual block. The filtering-target region and the filter type are both "determined based on the intra prediction mode of the current block."
• Dependent features: the filtering-target region comprises a left vertical prediction pixel line (leftmost column) and an upper horizontal prediction pixel line (top row); filtering is applied only for luma (not chroma); a fixed filter type is used regardless of block size; a 3-tap filter [1/4, 2/4, 1/4] at the top-left corner prediction pixel (using the top and left reconstructed neighbors), a horizontal 2-tap filter [1/4, 3/4] on the left column, and a vertical 2-tap filter [1/4, 3/4] on the top row; and mode-dependent offset prediction for vertical/horizontal modes (the "second aspect" of the disclosure).

In short, the invention is: mode-dependent smoothing of the boundary row/column of an intra-predicted block using reconstructed neighboring pixels, with specific 2-tap and 3-tap coefficient sets.

2. Prior art identified with high confidence

(a) ITU-T H.264 / ISO/IEC 14496-10 (AVC) — intra-prediction boundary smoothing (high confidence)

H.264, published well before the June 20, 2011 priority date, already performs mode-dependent filtering of intra prediction blocks using reconstructed neighboring pixels:

• For Intra_16x16_DC (luma) and 8x8 chroma DC, the DC value is used for the block interior, while the top row and left column are filtered/averaged with the adjacent reconstructed top and left reference pixels (2-tap averaging).
• H.264 also includes "constrained intra prediction" (CIP), which the patent itself discusses at length (Figs. 11–12), confirming that H.264 is part of the admitted background.

This alone teaches: intra-prediction mode (DC) → filtering target region = top row + left column → filter using reconstructed neighboring pixels → improved prediction and reduced boundary discontinuity.

(b) JCT-VC / HEVC development documents (high confidence that these existed; medium confidence on exact document numbers)

By the priority date, the HEVC standardization effort had already produced public proposals and working-draft text covering exactly this feature space:

• JCTVC-A119 ("Intra Prediction Improvements," Samsung, JCT-VC 1st meeting, Dresden, April 2010) is, to my knowledge, the well-known proposal that introduced DC-mode boundary filtering (3-tap [1/4, 2/4, 1/4] at the corner and 2-tap [1/4, 3/4] along the first row/column) and the planar mode into the HEVC test model.
• The HEVC Test Model (HM) working drafts circulating in late 2010 and the first half of 2011 (i.e., before the priority date) contained the DC boundary filter whose coefficient structure matches the patent's claimed 3-tap corner filter and 2-tap edge filters.

I flag that I cannot verify from the provided materials that any specific JCTVC document was cited by the Examiner; the identification of JCTVC-A119 is from my technical knowledge of the standard's history and should be confirmed against the file wrapper.

(c) The patent's own admitted background (high confidence)

The specification itself acknowledges: "the number of intra prediction modes ... may be 3, 5, 9, 17, 34, 35, 36" (i.e., H.264's 9 modes and HEVC's 34/35 modes), and it describes the standard reference-pixel geometry (below-left, left, above-left, above, above-right) and the DC/planar/directional mode set from the HEVC context. This confirms the invention operates squarely within the known H.264/HEVC intra-prediction framework.

3. Obviousness combinations and rationale

Combination A — H.264 (or JCT-VC/HM DC boundary filtering) + ordinary skill, generalized to all intra modes

Primary reference: H.264/AVC DC-mode boundary smoothing (top row + left column filtered with reconstructed neighbors).

What the claims add over H.264: (1) applying the same kind of boundary filtering not just to DC but to other intra modes (vertical, horizontal, diagonal modes), with the filter type and target region selected per mode; (2) the specific coefficient sets [1/4, 2/4, 1/4] and [1/4, 3/4].

Why obvious:

• The patent's own specification states the motivation: prediction pixels adjacent to reference pixels not used for a given directional mode have low correlation and high prediction error (e.g., left column pixels in a vertical mode, top row pixels in a horizontal mode). That is a recognized, well-documented problem in H.264/HEVC intra prediction.
• A PHOSITA facing that known problem would naturally extend the known DC-mode boundary-filtering solution (already in H.264) to the modes exhibiting the same boundary-discontinuity problem. Mode-dependent selection of filter region (left column for vertical-type modes; top row for horizontal-type modes) follows directly from the geometry of the prediction directions — it is the same boundary-error logic applied per mode.
• The coefficient values [1/4, 3/4] and [1/4, 2/4, 1/4] are simple, conventional smoothing weights (weighted averages using powers of two for cheap shift-based implementation). Selecting these weights is a routine design choice, not an inventive step.
• The luma-only limitation mirrors the known practice in H.264/HEVC of applying the more elaborate intra tools to luma while treating chroma more simply.

Conclusion for Combination A: Claim 1 and its dependent method/apparatus claims (3-tap corner filter, horizontal 2-tap, vertical 2-tap, luma-only, fixed filter type) would be obvious over H.264 in view of the knowledge of a PHOSITA, and a fortiori over JCT-VC/HM texts that already disclosed the exact DC boundary filter coefficients.

Combination B — JCT-VC/HM DC boundary filter + mode-dependent application from H.264 directional modes

Primary reference: JCT-VC/HM working draft disclosing the DC boundary filter with the exact claimed coefficients ([1/4, 2/4, 1/4] corner; [1/4, 3/4] row/column).

Secondary reference: H.264's directional intra modes and the known correlation problem at prediction-block boundaries.

Why obvious: If the prior-art DC boundary filter already teaches the identical filter structure and coefficients for the DC mode, then applying that same filter to the first row/column of other modes (vertical, horizontal, diagonal) is an obvious extrapolation: the boundary-error phenomenon is mode-dependent only in which boundary is affected, and the patent itself describes the mapping (left column for vertical-type modes, top row for horizontal-type modes). The "filtering target region and filter type determined based on the intra prediction mode" limitation is thus a direct, predictable application of the known DC filter to other known modes.

Combination C — the "second aspect" (offset-based vertical/horizontal prediction) over H.264/JCT-VC directional prediction

The claims directed to vertical-mode prediction using a first offset derived from left-neighbor differences, and horizontal-mode prediction using a second offset derived from top-neighbor differences, are essentially gradient/extrapolation-based refinements of H.264-style directional prediction. Directional intra prediction that extrapolates from neighboring reconstructed pixels along a direction was well known; adding a small offset derived from neighboring pixel differences (i.e., a local gradient term) is a standard least-squares/gradient-prediction technique that was explored extensively in the intra-prediction literature before 2011. A PHOSITA would be motivated to combine a known gradient predictor with the known H.264 directional modes to reduce DC-offset error, and the specific "left column for vertical mode / top row for horizontal mode" mapping is again dictated by the prediction geometry.

4. Graham-factor assessment

• Scope and content of prior art: H.264/AVC (boundary smoothing of DC intra blocks, CIP, directional modes) and JCT-VC/HM proposals (DC boundary filter with the claimed coefficients, planar mode, 35-mode set) are squarely in the same field and were publicly available before June 20, 2011.
• Differences between claims and prior art: The differences are (i) applying boundary filtering to modes beyond DC, (ii) mode-dependent selection of the boundary (left column vs. top row), and (iii) specific coefficient values. Each is a predictable, geometrically-driven variation of known techniques.
• Level of ordinary skill: A PHOSITA would be a video-coding engineer familiar with H.264 and the JCT-VC HEVC test model — precisely the population contributing to the HEVC standardization effort in 2010–2011, where these exact ideas (DC boundary filtering, mode-dependent smoothing) were being discussed and adopted.
• Secondary considerations: I have no evidence in the provided materials of long-felt need, unexpected results, licensing success, or industry praise that would rebut obviousness. The claimed coefficients are standard smoothing weights; no unexpected technical effect is apparent from the specification.

5. Caveats

• I could not confirm the Examiner's actual cited-reference list because the Prior Art section was not included in the supplied text and was not retrievable in my searches; the analysis above is built on references I can identify independently (H.264/AVC, JCT-VC/HM history). If the file wrapper cites additional references (e.g., Samsung/SK Telecom intra-filtering patent applications), those would only strengthen the obviousness case.
• JCTVC-A119 and the exact HM working-draft version should be verified against the JCT-VC document archive before being relied upon in litigation; my identification of the document number is from general knowledge of HEVC history, not from a confirmed citation of record.

6. Bottom line

Under 35 U.S.C. § 103, the claims of US 9332262 B2 are vulnerable to obviousness challenges based on (1) H.264/AVC's known DC-mode intra boundary filtering, and (2) JCT-VC/HM disclosures of the identical DC boundary filter coefficients, each in combination with the ordinary skill and well-known directional-intra geometry. The inventive contribution — extending known boundary smoothing to other intra modes with mode-dependent target regions and conventional [1/4, 3/4] / [1/4, 2/4, 1/4] weights — is a predictable variation that a PHOSITA would have had ample motivation to make.