PTAB

IPR2024-01503

Amazon.com Inc v. Nokia Technologies Oy

Log In
Key Events
Petition
petition Intelligence

1. Case Identification

2. Patent Overview

  • Title: Method for Reducing Redundancy in Video Coding Motion Vector Prediction
  • Brief Description: The ’833 patent discloses a method for improving video compression efficiency within video coding standards like H.265/HEVC. The invention focuses on reducing redundant candidates in a "merge list" used for motion vector prediction by performing a limited number of comparisons between a potential candidate and a subset of existing candidates, rather than comparing every available candidate pair.

3. Grounds for Unpatentability

Ground 1: Obviousness over Rusert and Zheng - Claims 9-14, 16, 18, and 26-30 are obvious over Rusert in view of Zheng.

  • Prior Art Relied Upon: Rusert (Application # 2011/0194609) and Zheng (Application # 2013/0077691).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued Rusert teaches the core inventive concept in the context of the H.264 video coding standard. Rusert disclosed a method to de-duplicate a list of predicted motion vector (PMV) candidates by comparing a new potential candidate only against the subset of candidates already added to the list. This avoids comparing every possible pair of candidates, thereby reducing computational complexity. Petitioner asserted that Zheng provided the necessary teachings to adapt Rusert's H.264-based method to the then-emerging H.265 standard. Zheng described the analogous H.265 concepts of "prediction units" (PUs) and "merge lists," making the application of Rusert's logic straightforward. The combination allegedly taught every limitation of independent claim 9, including determining a subset of candidates based on the current block's location and comparing a new candidate to that subset to decide on its inclusion in the merge list.
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine these references because Rusert explicitly stated its principles could be applied to other coding standards. H.265 was the well-known successor to H.264, and both standards shared a fundamental block-based motion prediction architecture. A POSITA seeking to improve H.265 efficiency, as taught by Zheng, would have been motivated to apply known optimization techniques from H.264, such as the candidate list de-duplication method taught by Rusert.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success. The combination represented a simple application of a known de-duplication technique (Rusert) to an analogous system (Zheng's H.265 framework). The predictable result would be improved coding efficiency by reducing redundant candidates, the stated goal of both references.

Ground 2: Obviousness over Nakamura and WD4 - Claims 9-14, 16, 18, and 26-30 are obvious over Nakamura in view of WD4.

  • Prior Art Relied Upon: Nakamura (JCTVC-F419, a July 2011 H.265 standards proposal) and WD4 (JCTVC-F803, an H.265 working draft from the same standards meeting).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner contended that Nakamura, a formal proposal to the H.265 standards body, taught a specific method to reduce comparisons when building a merge list. Nakamura proposed deriving two spatial merging candidates, S0 and S1, from a set of five possible spatial neighbors. Crucially, to remove redundancy, the motion information of S1 was compared only with the motion information of S0. Petitioner argued this directly taught comparing a candidate against a defined subset (here, a single other candidate) without comparing all possible pairs, thus meeting the core limitation of claim 9. WD4, the relevant H.265 working draft at the time, provided the necessary technical context, defining the merge list (mergeCandList) and prediction unit structure that Nakamura's proposal was designed to improve.
    • Motivation to Combine: The motivation to combine was explicit and direct. Nakamura was submitted as a proposed improvement to the H.265 standard, which was embodied in the WD4 working draft. A POSITA involved in the standards development process would have been directly motivated to apply Nakamura's proposed simplification to the WD4 draft to evaluate its merits, as this was the intended purpose of the proposal.
    • Expectation of Success: There was a very high expectation of success. Nakamura was specifically designed to be implemented into the H.265 test model software and the proposal itself included data showing performance gains from the combination. This demonstrated that the proposed technique was not only compatible with the existing draft but yielded the desired predictable improvement.

4. Key Claim Construction Positions

  • “the block”: Petitioner argued that, consistent with the prosecution history, this term in limitation [9d] referred to the "block of pixels" (i.e., the current block being processed) introduced in limitation [9a]. This construction was important because it anchored the determination of the "subset of... candidates" to the location of the current block.
  • “a subset of… candidates”: Petitioner asserted this phrase meant a subset of one or more candidates. This construction allowed a comparison against just one other candidate (as taught by Nakamura) to satisfy the claim limitation, which required comparing a candidate against "another... candidate of the [sub]set."

5. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under §314(a) based on Fintiv was unwarranted because the parallel district court case was in its nascent stages, with no case management schedule issued.
  • Petitioner further argued that denial under §325(d) was inappropriate because the core prior art and arguments were never considered by the USPTO. The Examiner's reason for allowing the ’833 patent was the "reduced-comparison" limitation. However, Rusert and Zheng (Ground 1) were never before the Examiner. While Nakamura (part of Ground 2) was cited in an IDS, there was no evidence the Examiner considered the full reference or, critically, the combination of Nakamura with WD4, which Petitioner argued was highly material and taught the exact feature upon which the patent was granted.

6. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 9-14, 16, 18, and 26-30 of the ’833 patent as unpatentable.