PTAB
IPR2018-00696
Samsung Electronics Co Ltd v. M & K Holdings Inc
Key Events
Petition
Table of Contents
petition
1. Case Identification
- Patent #: 9,113,163
- Filed: February 26, 2018
- Petitioner(s): Samsung Electronics Co., Ltd.
- Patent Owner(s): M & K Holdings Inc.
- Challenged Claims: 1-6
2. Patent Overview
- Title: Method of Decoding a Moving Picture in Inter Prediction Mode
- Brief Description: The ’163 patent discloses a method for decoding a video's motion vectors in inter prediction mode. The method seeks to improve coding efficiency by constructing a motion vector candidate list using motion vectors from spatially and temporally neighboring prediction units, which is then used to predict the motion vector of a current block.
3. Grounds for Unpatentability
Ground 1: Anticipation of Claims 1, 2, 5, and 6 - Claims 1, 2, 5, and 6 are anticipated under 35 U.S.C. §102 by WD4-v3.
- Prior Art Relied Upon: WD4-v3 (JCTVC-F803 (version 3), a working draft of the High-Efficiency Video Coding (HEVC) standard).
- Core Argument for this Ground:
- Prior Art Mapping: Petitioner contended that WD4-v3, a foundational document for the HEVC standard, discloses every limitation of independent claim 1 and its dependent claims 2, 5, and 6. Petitioner mapped the claim elements to WD4-v3’s disclosed decoding process, which includes generating prediction and residual blocks and deriving motion vector components. Specifically for claim 1(j), Petitioner argued WD4-v3 discloses the “first available motion vector” alternative by describing a process that checks for a temporal motion vector candidate first from a right-bottom block and then from a center block, both corresponding to the current prediction unit.
- Key Aspects: Petitioner asserted that WD4-v3’s detailed description of the HEVC decoding process, including entropy-decoding, inverse scanning, and motion vector candidate list construction (including adding a zero vector when the candidate list is not full), constitutes a complete disclosure of the claimed invention.
Ground 2: Obviousness of Claim 2 - Claim 2 is obvious over WD4-v3 in view of Park.
- Prior Art Relied Upon: WD4-v3 (JCTVC-F803 (version 3)) and Park (JCTVC-E059 (version 4), a contribution to the JCT-VC).
- Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that while WD4-v3 teaches using an upper left pixel of the center candidate block (the first alternative of claim 2(c)), it does not explicitly teach using a below right pixel (the second alternative). The Park reference, a proposal to improve the HEVC standard, teaches modifying the temporal motion vector predictor to use a position corresponding to a below right pixel of the center block.
- Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Park with the WD4-v3 framework because Park was a formal contribution to the HEVC standardization process specifically intended to improve the methods being developed in working drafts like WD4-v3. Park’s explicit purpose was to improve coding efficiency, a primary goal of the HEVC standard.
- Expectation of Success: Park expressly stated that its proposed change achieves a “meaningful coding gain... without noticeable impact on complexity.” This statement would have provided a POSITA with a strong reason to believe the combination would be successful and yield a predictable improvement.
Ground 3: Obviousness of Claims 3-4 - Claims 3-4 are obvious over WD4-v3 and Park, in view of Zhou.
- Prior Art Relied Upon: WD4-v3 (JCTVC-F803 (version 3)), Park (JCTVC-E059 (version 4)), and Zhou (JCTVC-G082 (version 1), a contribution to the JCT-VC).
- Core Argument for this Ground:
- Prior Art Mapping: This ground built upon the combination of WD4-v3 and Park. It addressed the alternative limitation in claim 1(j) where the temporal motion vector candidate is the “motion vector of a predetermined block,” which is required by claim 3. Petitioner argued the Zhou reference teaches a method where, if the current prediction unit is located at a lower boundary of a Largest Coding Unit (LCU), the temporal motion vector is selected from a specific candidate block (C3). This selection rule makes the candidate block "predetermined" under the condition specified in claim 4.
- Motivation to Combine: A POSITA would have been motivated to incorporate Zhou’s logic into the WD4-v3/Park process to achieve Zhou’s stated goal of reducing memory bandwidth for temporal motion vector derivation. Like Park, Zhou was a proposal submitted to improve the HEVC standard, making its teachings directly relevant to a POSITA working with WD4-v3.
- Expectation of Success: The proposed change in Zhou was straightforward to implement and offered a clear benefit in memory efficiency. The fact that Zhou’s configuration was later incorporated into the next working draft of the HEVC standard would have confirmed to a POSITA that the combination was successful.
4. Key Claim Construction Positions
- "below left pixel" (Claim 2): Petitioner argued this term is a typographical error and should be construed as “below right pixel.” This assertion was based on an inconsistency between the claim text and Figure 5 of the ’163 patent, which clearly illustrates a below right pixel. This interpretation is further supported by the specification of the priority PCT application and claims of other patents in the same family, which all consistently refer to a "below right pixel" in the same context. This construction is critical to the obviousness arguments in Grounds 2 and 3.
- "if..." clauses (Claims 4 and 6): Petitioner contended that the conditional limitations in claims 4 and 6 (e.g., “if a size of the transform unit is larger than a first reference size...”) should not be given patentable weight because they frame the feature as optional. However, Petitioner also argued in the alternative that, should the Board give the limitations weight, the prior art discloses the conditions and the resulting actions.
5. Key Technical Contentions (Beyond Claim Construction)
- Prior Art Status as "Printed Publications": A central contention of the petition was that the prior art references (WD4-v3, Park, and Zhou) qualify as printed publications under §102. Petitioner argued these JCT-VC documents were made publicly available before the critical date. This was achieved through dissemination on the JCT-VC document management site, which was publicly accessible without restriction, and by being mirrored on the MPEG site. While the MPEG site required credentials, Petitioner provided evidence that these credentials were systematically and widely distributed to hundreds of members, making the documents accessible to those interested and skilled in the art exercising reasonable diligence.
6. Relief Requested
- Petitioner requests institution of an inter partes review and cancellation of claims 1-6 of Patent 9,113,163 as unpatentable.
Analysis metadata