Amazon.com Inc v. Nokia Technologies Oy

1. Case Identification

• Case #: IPR2024-00626
• Patent #: 11,805,267
• Filed: March 15, 2024
• Petitioner(s): Amazon.com, Inc., Amazon.com Services LLC
• Patent Owner(s): Nokia Technologies Oy
• Challenged Claims: 1-18

2. Patent Overview

• Title: Utilizing Motion Prediction in Video Coding
• Brief Description: The ’267 patent describes a method for reducing rounding errors in bi-directional motion prediction for video encoding. The method involves calculating prediction signals using a higher precision (more bits) for intermediate steps and then rounding the combined prediction down to a lower precision for the final output.

3. Grounds for Unpatentability

Ground 1: Obviousness over Walker - Claims 1-18 are obvious over Walker

• Prior Art Relied Upon: Walker (Application # 2005/0281334).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Walker taught a weighted prediction method for H.264 video coding that anticipated every limitation of the challenged claims. Independent claims 1, 7, and 13 require using a "first precision" for pixel values and a "second precision, which is higher than said first precision," for intermediate predictions before "decreasing a precision" of the final combined prediction. Petitioner asserted Walker’s Table 2 explicitly disclosed this process by showing initial 8-bit pixel values (first precision) being used to calculate 16-bit intermediate predictions (higher second precision). Walker then showed this combined prediction's precision being decreased from 18 bits to 10 bits via a right bit-shift operation, directly mapping to the claimed "decreasing" step.
  • Motivation to Combine (for §103 grounds): Not applicable as this ground is based on a single reference. Petitioner argued Walker alone taught all elements.
  • Expectation of Success (for §103 grounds): Not applicable.
  • Key Aspects: Petitioner contended that Walker’s detailed tables of bit-widths for each operational step provided an explicit roadmap for a person of ordinary skill in the art (POSITA) to arrive at the claimed invention. Dependent claims related to sub-pixel interpolation, rounding offsets, and bit-shifting were also argued to be expressly taught or made obvious by Walker's disclosure of standard H.264-related video processing techniques.

Ground 2: Obviousness over Karczewicz-I and Karczewicz-II - Claims 1-18 are obvious over Karczewicz-I in view of Karczewicz-II

• Prior Art Relied Upon: Karczewicz-I (Application # 2011/0007799) and Karczewicz-II (Application # 2009/0257499).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued Karczewicz-I taught the framework for bi-directional prediction in H.264 video coding, including averaging two predictions from reference blocks. Karczewicz-II addressed the known problem of rounding inaccuracies in pixel interpolation by teaching a method to "keep the highest possible precision through the intermediate steps" and avoiding "any shifting, rounding and clipping operations until the very last step." Petitioner asserted Karczewicz-II explicitly taught using higher-precision, non-rounded intermediate values (e.g., 15-bit half-pixel predictions) and then combining them before a final right-shift operation reduces the precision. This combination directly taught the core inventive concept of the ’267 patent.
  • Motivation to Combine (for §103 grounds): A POSITA would combine these references to improve the accuracy of the bi-prediction process taught in Karczewicz-I. Karczewicz-II provided a known solution—maintaining higher intermediate precision—to address the well-understood problem of compounding rounding errors inherent in the calculations of Karczewicz-I. The references were from the same inventors and addressed complementary aspects of the same H.264 video coding architecture, providing a strong motivation to integrate their teachings.
  • Expectation of Success (for §103 grounds): A POSITA would have a high expectation of success because the combination applied a known mathematical technique (delaying rounding) to a standard video coding process to achieve the predictable result of improved prediction accuracy.

4. Key Claim Construction Positions

• "precision": Petitioner argued that a POSITA would have understood "precision" to be satisfied by, but not limited to, "a number of bits needed to represent possible values." This construction was asserted to be consistent with the specification, which describes precision in terms of bit-depth (e.g., N-bit pixels, M-bit predictions), and the prosecution history. Defining precision this way was critical to comparing the "first precision" of pixels with the "higher" "second precision" of intermediate predictions as required by the claims.

5. Arguments Regarding Discretionary Denial

• Fintiv: Petitioner argued discretionary denial under Fintiv was not warranted. Since the ’267 patent was asserted in a parallel ITC proceeding, the related district court proceedings are subject to a mandatory stay. Petitioner cited PTAB policy stating it will not discretionarily deny petitions based on Fintiv in such circumstances.
• Advanced Bionics (§325(d)): Petitioner argued denial was not warranted because the primary references were not meaningfully considered during prosecution. Walker and Karczewicz-I were never cited. Although Karczewicz-II was cited in an Information Disclosure Statement (IDS), it was one of 77 references in that IDS and the Examiner performed no substantive analysis on it.

6. Relief Requested

• Petitioner requested the institution of an inter partes review and the cancellation of claims 1-18 of the ’267 patent as unpatentable.