PTAB

IPR2015-00348

Google Inc v. Network 1 Technologies Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Algorithms for Automatic Content Recognition
  • Brief Description: The ’441 patent describes a system for identifying an unknown electronic work (e.g., audio or video) and performing a corresponding action. The system operates by extracting features from the unknown work, performing a non-exhaustive neighbor search against a database containing data for reference works, and using the identification to determine and execute an action associated with the reference work.

3. Grounds for Unpatentability

Ground 1: Obviousness over Levy and Arya - Claims 1-3, 6-14, 18-19, 21-26, and 30 are obvious over Levy in view of Arya.

  • Prior Art Relied Upon: Levy (Patent 6,505,160) and Arya (a 1998 article titled "An Optimal Algorithm for Approximate Nearest Neighbor Searching in Fixed Dimensions").
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Levy discloses a system for identifying a media work and executing a related action. Levy's system extracts a unique identifier or "fingerprint" from an audio signal, transmits it to a server, and uses it to look up a corresponding action in a database. However, Levy's primary system relies on finding an exact match. Petitioner contended that Arya discloses an efficient, non-exhaustive neighbor search algorithm designed to find "approximate nearest neighbors" and explicitly mentions its applicability to multimedia databases. The combination of Levy's content identification framework with Arya's efficient near-match search algorithm renders the challenged claims obvious.
    • Motivation to Combine: A POSITA would combine Levy with Arya to improve Levy's functionality. Petitioner asserted that a POSITA would have understood that content fingerprinting algorithms like those in Levy could generate slightly different fingerprints from disparate sources of the same work. This would necessitate searching for not only exact matches but also near matches. Arya provides a well-known, efficient, and non-exhaustive algorithm to solve this exact problem, making it an obvious modification to improve the robustness of Levy's system.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success in combining the references. The integration involved applying a known search algorithm (Arya) to a known content identification system (Levy), a task well within the ordinary skill in the art.

Ground 2: Anticipation by Conwell - Claims 1-3, 6, 8-14, 19, 21-26, and 30 are anticipated by Conwell.

  • Prior Art Relied Upon: Conwell (Patent 6,970,886).

  • Core Argument for this Ground:

    • Prior Art Mapping: Petitioner argued that Conwell discloses every element of the independent claims. Conwell describes a method for directing users of digital content to related websites. The system extracts an identifier from the content by applying a hashing algorithm, stores these identifiers in a database with corresponding web addresses (the "action"), and uses a user-generated identifier to look up the URL. Petitioner contended that Conwell's use of a hashing algorithm—which can map similar but non-identical inputs to the same hash output—inherently performs a "neighbor search." Furthermore, the direct table look-up method is a "non-exhaustive search" because it avoids a brute-force comparison of all possible matches. Conwell's disclosure of a database linking content identifiers to actions (URLs) was argued to meet the remaining claim limitations.

  • Additional Grounds: Petitioner asserted additional challenges including that claims 1-3, 8-14, 19, 21-22, 25, 27, and 29-30 are anticipated by Wood (Patent 7,743,092), and that claims 1-3, 8, 10-14, 18-19, 21-27, and 29-30 are obvious over Ghias (Patent 5,874,686) in view of Philyaw (Patent 6,098,106).

4. Key Claim Construction Positions

Petitioner proposed constructions for several key terms that are central to its unpatentability arguments, asserting they should be given their broadest reasonable construction.

  • "extracted features" / "compact electronic representation": Petitioner argued these synonymous terms should be construed as "a quantitative representation of the features of a media work, or a subset of the features of a media work." This broad construction is supported by the specification and encompasses various known feature extraction methods, including fingerprinting and hashing, as taught in the prior art.
  • "non-exhaustive...search": Proposed construction is "a search that locates a match without conducting a brute force comparison of all possible matches, and all data within all possible matches." This construction is critical for arguing that the efficient look-up tables and approximate matching algorithms in the prior art meet this limitation.
  • "neighbor search": Proposed construction is "identifying a close, but not necessarily exact, match." This construction allows prior art that finds approximate or functionally similar matches (like through hashing in Conwell or distance scores in Wood) to meet the claim limitation.

5. Relief Requested

  • Petitioner requests the institution of an inter partes review and the cancellation of claims 1-3, 6-14, 18-19, 21-27, and 29-30 of the ’441 patent as unpatentable under 35 U.S.C. §§ 102 and 103.