PTAB

IPR2015-00345

Google Inc v. Network 1 Technologies Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Automatic Content Recognition System
  • Brief Description: The ’237 patent describes computer-implemented methods and systems for identifying a media work. The technology involves receiving features extracted from a media work, using those features to search a database to determine the work's identity, and then either transmitting information about the work or determining an action based on the identification.

3. Grounds for Unpatentability

Ground 1: Anticipation over Iwamura - Claims 1, 3-5, 7-9, 11-13, 15-16, 21-25, 29-30, 33, and 37-38 are anticipated by Iwamura.

  • Prior Art Relied Upon: Iwamura (Patent 6,188,010).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Iwamura, which discloses a system for identifying a melody input by a user, teaches every element of the challenged claims. Iwamura discloses receiving features extracted from a melody (e.g., "peak notes" or features derived from a Fast Fourier Transform). It then performs a search of a remote music database to find the "closest melody," which Petitioner contended meets the "neighbor" search limitations. Finally, Iwamura performs an action by sending back search results, providing a link to an online music store for purchase, or sending a MIDI file of the song to the user. For claims requiring a "nonexhaustive search," Petitioner asserted that Iwamura's method of searching using only "peak notes"—which constitute approximately 20% of a melody's total notes—is inherently nonexhaustive. For claims requiring a "sublinear" search, Petitioner argued that Iwamura’s disclosure of using the Boyer-Moore algorithm constitutes a sublinear search.

Ground 2: Obviousness over Levy in view of Arya - Claims 1-25, 29-30, 32-33, 37-38, and 40 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 asserted that Levy taught the foundational system of identifying media and executing a related action. Levy discloses generating a "fingerprint" (i.e., extracted features) from an audio signal, looking up that fingerprint in a database to identify the media object, and then performing an action like displaying the title/artist or facilitating an electronic transaction. Petitioner argued that Levy lacked an explicit disclosure of an efficient, approximate, or sublinear search. This missing element, however, was supplied by Arya, which discloses efficient algorithms (e.g., using kd-trees) for performing approximate nearest neighbor searches that are sublinear and inherently nonexhaustive.
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would have understood that fingerprinting algorithms can produce slight variations, making it necessary to search for near matches, not just exact ones. Petitioner argued a POSITA would combine Levy’s system with Arya’s well-known efficient search algorithm to improve the system's performance and handle such variations, especially since Arya explicitly mentions applicability to multimedia databases.
    • Expectation of Success: The combination was portrayed as the predictable application of a known, efficient search technique (Arya) to a known content identification system (Levy) to achieve improved speed and functionality.

Ground 3: Obviousness over Iggulden in view of Böhm - Claims 25, 32-33, and 40 are obvious over Iggulden in view of Böhm.

  • Prior Art Relied Upon: Iggulden (Patent 6,597,405) and Böhm (a 2000 article titled "Efficient Similarity Search in Digital Libraries").
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner contended that Iggulden taught a system for identifying known television commercials and performing a subsequent action, such as muting the TV. Iggulden extracts a "signature pattern" (features) from a television segment and compares it to stored patterns to find a match. Petitioner argued Böhm supplied the claimed nonexhaustive and sublinear search methods. Böhm discloses high-performance, index-based search algorithms for finding similar objects in complex data, including k-nearest neighbor searches that are both nonexhaustive (by considering only a fraction of feature vectors) and sublinear.
    • Motivation to Combine: A POSITA would have been motivated to implement the search in Iggulden’s system using a more efficient algorithm like that taught by Böhm. Böhm’s algorithm was described as outperforming other scan-based methods and was shown to be applicable to relevant data types, including "TV-snapshots." This made it a logical choice for improving the performance of Iggulden's commercial detection system.
    • Expectation of Success: Applying Böhm’s advanced search methodology to Iggulden’s system to increase search efficiency would have been a predictable design choice with a high expectation of success.
  • Additional Grounds: Petitioner asserted additional anticipation challenges against various claims based on Ghias (Patent 5,874,686) and Wood (Patent 7,743,092). Petitioner also asserted obviousness challenges that added Chen (Patent 7,444,353) to the Iwamura and Levy/Arya combinations to teach transmitting information to a separate device.

4. Key Claim Construction Positions

  • "features extracted from the media work": Petitioner proposed this term be construed as "a quantitative representation of the features of a media work, or a subset of the features of a media work," arguing this is the only unifying characteristic of the various extraction methods disclosed.
  • "nonexhaustive search": Proposed construction was "a search that locates a match without conducting a brute force comparison of all possible matches, and all data within all possible matches."
  • "identify a neighbor" / "nearest neighbor search": These related terms were proposed to mean "identify a close, but not necessarily exact, match," consistent with the common understanding in the art at the time.
  • "sublinear search": Proposed construction was "a search whose execution time has a sublinear relationship to database size," meaning its execution time does not grow proportionally with the size of the database.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-27, 29-30, 32-35, 37-38, and 40 of the ’237 patent as unpatentable.