PTAB

IPR2024-00847

Amazon.com Inc v. Nokia Technologies Oy

Log In
Key Events
Petition
petition Intelligence

1. Case Identification

2. Patent Overview

  • Title: Method for Coding Motion in a Video Sequence
  • Brief Description: The ’808 patent discloses methods for encoding and decoding video sequences by improving the efficiency of video compression. The invention purports to redefine the "skip mode" concept to allow a macroblock to be associated with either a zero motion vector (as in conventional skip modes) or a predicted non-zero motion vector based on the motion of neighboring segments, thereby reducing the amount of data needed to represent motion.

3. Grounds for Unpatentability

Ground 1: Obviousness over Karczewicz - Claims 1-2, 6-7, 9, 23-25, 28-30, 32, 34, 36, 39-41, 43, and 65 are obvious over Karczewicz.

  • Prior Art Relied Upon: Karczewicz (WO 01/11891 A1).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Karczewicz teaches a method for video encoding with several distinct coding modes, including modes that predict a segment’s motion field from a neighboring segment (resulting in a non-zero motion vector) and a mode that uses a zero motion vector (ZMV) field. Petitioner contended that these two functionalities, taken together, correspond to the ’808 patent’s redefined "skip coding mode." Karczewicz’s use of a single Motion Coefficient Indication (MCI) bit to signal when no refinement field is transmitted (which occurs in both the ZMV and predicted modes) was argued to meet the limitation of "providing...an indication of the skip coding mode." The decision to use a particular mode is based on efficiency, which Petitioner mapped to the claim requirement of assigning a motion vector based on a neighboring segment's motion information.
    • Motivation to Combine (Implicit): This ground relies on a single reference. Petitioner argued a person of ordinary skill in the art (POSITA) would have understood that Karczewicz's distinct coding modes that utilize either a ZMV or a predicted motion vector collectively function as the claimed "skip coding mode," as both are signaled by a zero MCI bit and represent low-bitrate prediction options.
    • Expectation of Success: Not applicable for a single-reference ground.

Ground 2: Obviousness over Frojdh - Claims 1-2, 7, 23, 28-30, 34, 39-41, and 65 are obvious over Frojdh.

  • Prior Art Relied Upon: Frojdh (Application # 2003/0123738).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Frojdh discloses an Implicit Global Motion Compensation (IGMC) mode and a COPY (or SKIP) mode for video compression. The COPY mode uses a ZMV, while the IGMC mode uses a predicted non-zero motion vector (PNZMV) derived from neighboring blocks. Petitioner argued that these modes, particularly when available under Frojdh’s Global Motion Compensation (GMC) framework, teach the claimed skip mode that selectively uses either a ZMV or a PNZMV. Frojdh teaches that the choice between these modes depends on coding efficiency and motion characteristics, which satisfies the limitation of assigning the vector based on neighboring motion information. Frojdh’s IGMC and COPY modes also operate without explicitly coding motion vectors, meeting the "no further motion vector information" limitation.
    • Motivation to Combine (Implicit): This ground relies on a single reference. Petitioner argued a POSITA would recognize that Frojdh’s IGMC and COPY modes are complementary tools for efficient coding that, when enabled together, directly teach the claimed invention.
    • Expectation of Success: Not applicable for a single-reference ground.

Ground 3: Obviousness over Frojdh and H.263 - Claims 6, 9, 24-25, 32, 36, and 43 are obvious over Frojdh in view of H.263.

  • Prior Art Relied Upon: Frojdh (Application # 2003/0123738) and H.263 (ITU-T Recommendation H.263).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Frojdh teaches the high-level concept of predicting a motion vector from neighboring blocks in its IGMC mode but omits specific implementation details. H.263, a prominent video coding standard, explicitly teaches a method for calculating a motion vector predictor by taking the median of the motion vectors from three surrounding macroblocks. The combination, therefore, teaches the specific mechanism for deriving the PNZMV contemplated by Frojdh. Specifically, H.263’s rules for handling boundary conditions and INTRA-coded neighbors teach how a neighboring segment with a ZMV would lead to the assignment of a ZMV for the current segment, directly mapping to the limitations of claims 6, 9, and 24.
    • Motivation to Combine: A POSITA would combine the references because Frojdh presents IGMC as an improvement to existing standards like H.263 but does not provide a complete implementation. H.263 provides the well-known, standardized "off-the-shelf" solution for motion vector prediction needed to fully implement Frojdh’s conceptual teachings. A POSITA would have naturally looked to an established standard like H.263 to fill in the predictable details of Frojdh’s system.
    • Expectation of Success: Petitioner argued a POSITA would have a high expectation of success, as this combination involves applying a known, predictable technique (H.263’s predictor calculation) to a known problem (implementing Frojdh’s predictive IGMC mode), which would yield predictable results without altering the core functionality of either system.

4. Key Claim Construction Positions

  • Petitioner argued that the term "skip coding mode" should be construed to mean a coding mode where a macroblock is associated with either a zero or non-zero motion vector, and where residual information (prediction error) can be provided.
  • This construction is central to Petitioner’s arguments, particularly for the Karczewicz ground, as Karczewicz explicitly teaches sending prediction error with its coding modes.
  • Petitioner supported this position by invoking the doctrine of claim differentiation, pointing to dependent claims (e.g., claim 30) that explicitly recite "wherein no residual information is provided," which would be rendered superfluous if the base claim’s "skip coding mode" already excluded residual information.

5. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under Fintiv should not apply. It was noted that at the time of filing the petition, no trial date or substantive schedule had been set in the co-pending district court litigation, suggesting that an IPR would be a more efficient resolution.

6. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-2, 6-7, 9, 23-25, 28-30, 32, 34, 36, 39-41, 43, and 65 of Patent 7,532,808 as unpatentable.