PTAB

IPR2017-00100

SamSung ElecTronics Co Ltd v. InfobRidge Pte Ltd

Log In

1. Case Identification

  • Patent #: 8,917,772
  • Filed: October 17, 2016
  • Petitioner(s): Samsung Electronics Co., Ltd.
  • Patent Owner(s): Infobridge Pte. Ltd.
  • Challenged Claims: 8-9

2. Patent Overview

  • Title: Method of Constructing a Merge Candidate List in Video Coding
  • Brief Description: The ’772 patent relates to video coding technologies, specifically a method for constructing a merge candidate list for a current prediction unit during video encoding or decoding. The method uses spatial and temporal neighboring blocks to derive motion information, improving coding efficiency.

3. Key Technical Contentions (Beyond Claim Construction)

  • Ineffective Priority Claim: Petitioner argued that the ’772 patent is not entitled to the November 7, 2011 filing date of its Korean priority application (’219 KR application). This argument was based on two primary points:
    • There is no common inventorship between the ’772 patent (inventors Soo Mi Oh and Moonock Yang) and the final named inventor of the ’219 KR application (Min Jang), as required under §119(a).
    • The ’219 KR application does not provide sufficient written description support for the limitation in claim 9 that "a reference picture index of the temporal merge candidate is 0."
    • As a result, Petitioner contended the challenged claims have an effective filing date no earlier than November 7, 2012, making the asserted prior art references, all published before this date, valid under at least §102(b).
  • Prior Art as Printed Publications: Petitioner asserted that the Joint Collaborative Team on Video Coding (JCT-VC) documents relied upon as prior art (WD4, WD3, Zhou) qualify as "printed publications." The argument was that these documents were made publicly available on the JCT-VC document management website without restriction and were also mirrored on the MPEG website, accessible to hundreds of members. Further, announcements of their availability were disseminated via email reflectors to hundreds of interested parties in the video coding field.

4. Key Claim Construction Positions

  • Petitioner asserted that under the Broadest Reasonable Interpretation (BRI) standard, no terms in the challenged claims require construction. However, the petition preemptively addressed potential narrow constructions that the Patent Owner might propose, which formed the basis for several obviousness grounds:
    • Claim 8(b): For the limitation "a position of the temporal merge candidate block is determined depending on a position of the current block within an LCU," Petitioner argued against a narrow construction requiring the position to be determined from among available temporal merge candidate blocks.
    • Claim 9: For the limitation "a reference picture index of the temporal merge candidate is 0," Petitioner argued against a narrow construction requiring the index to be always set to 0.

5. Grounds for Unpatentability

Ground 1: Anticipation of Claims 8 and 9 - Claims 8 and 9 are anticipated by WD4.

  • Prior Art Relied Upon: WD4 (JCTVC-F803, version 4, a working draft of the HEVC standard, published Oct. 4, 2011).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that WD4 discloses every limitation of claims 1, 8, and 9. WD4 allegedly teaches a method of constructing a merge candidate list (mergeCandList) by checking the availability of spatial candidates (A0, A1, B0, B1, B2) and a temporal candidate (Col), and using their motion information. It also describes setting a spatial merge candidate as unavailable if the current block is the second of two asymmetrically partitioned units. For claim 8, WD4 teaches determining the position of the temporal candidate block based on the position of the current block within a Largest Coding Unit (LCU). For claim 9, WD4 teaches setting the reference picture index of the temporal merge candidate to 0 under certain conditions.

Ground 2: Obviousness of Claim 8 - Claim 8 is obvious over WD4 in view of Zhou.

  • Prior Art Relied Upon: WD4, and Zhou (JCTVC-F081, version 2, an evaluation of an alternative HEVC solution, published July 11, 2011).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground was presented as an alternative to Ground 1 should the Patent Owner successfully argue for a narrower construction of claim 8(b), requiring that the temporal merge candidate block be selected from among available candidates based on the current block's position. While WD4 teaches determining the candidate's position, Zhou explicitly discloses a process where one of two potential temporal candidates ("H" and "BR") is selected based on whether the current prediction unit is adjacent to a lower LCU boundary.
    • Motivation to Combine: A POSITA would combine WD4 with Zhou because both references address the same HEVC standard development and aim to improve video coding efficiency. Zhou's method of selecting from available candidates based on block position was proposed to reduce the memory bandwidth required to fetch co-located motion data, offering a clear benefit and predictable solution to a known problem in the art.
    • Expectation of Success: A POSITA would have had a high expectation of success, as Zhou's technique was a specific, targeted improvement for the temporal merge process already defined in working drafts like WD4.

Ground 4: Obviousness of Claims 8 and 9 - Claims 8 and 9 are obvious over WD4 in view of Han.

  • Prior Art Relied Upon: WD4, and Han (an IEEE Transactions article on flexible unit representation, published Dec. 2010).

  • Core Argument for this Ground:

    • Prior Art Mapping: This ground addressed the potential argument that WD4 does not explicitly define certain partition modes (e.g., Part_nLx2N) as "asymmetric partitioning," as required by claim 1(h). Han was introduced because it explicitly discloses and defines these exact same partition modes as "asymmetric PU splittings" designed to improve coding efficiency for irregular object boundaries.
    • Motivation to Combine: A POSITA would combine WD4 with Han to formally define the partition modes in WD4 as asymmetric. Both references relate to HEVC development, and Han's disclosure provides the precise terminology and rationale for using such partitions, which was to overcome restrictions imposed by symmetric-only partitions. This would have been a common-sense implementation to clarify and enhance the functionality described in WD4.
    • Expectation of Success: Success would be expected because the combination merely involves applying the known definition and benefits from Han to the specific, existing partition structures in WD4, without requiring substantial modification.

  • Additional Grounds: Petitioner asserted additional obviousness challenges, including that claim 9 is obvious over WD4, Zhou, and WD3 (an earlier HEVC working draft) if the reference picture index must always be set to 0. Further grounds combined WD4, Han, and Zhou (for claim 8) and WD4, Han, Zhou, and WD3 (for claim 9), building upon the same core arguments.

6. Relief Requested

  • Petitioner requests institution of inter partes review and cancellation of claims 8 and 9 of the ’772 patent as unpatentable based on all asserted grounds.