PTAB

IPR2020-00052

Netflix Inc v. DivX LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Method of Deblocking Reconstructed Video Frame
  • Brief Description: The ’651 patent relates to a method for reducing blocking artifacts in compressed digital video. The method involves analyzing a two-dimensional region of pixels around a block boundary to determine a "level of detail" and then selecting an appropriate deblocking filter based on that determination.

3. Grounds for Unpatentability

Ground 1: Anticipation and Obviousness over Vehvilainen - Claims 1 and 17-19 are anticipated or obvious over Vehvilainen.

  • Prior Art Relied Upon: Vehvilainen (Patent 6,504,873).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Vehvilainen taught every element of the challenged claims. Vehvilainen disclosed a method of deblocking video that first determined a "level of detail" by calculating pixel variance (termed "activityIN") within a two-dimensional region around a block boundary. This region, shown in Vehvilainen's Figure 5A, included pixels from multiple rows and columns on both sides of the boundary, with at least one pixel not immediately adjacent to the boundary, thus meeting the region requirements of claim 1. Based on whether the region was "smooth," "moderate," or "high activity," Vehvilainen selected a corresponding filter. This process of determining detail and selecting a filter based on that detail allegedly met all limitations of independent claim 1.
    • Prior Art Mapping (Dependent Claims): For claim 17, Petitioner asserted Vehvilainen taught comparing the determined level of detail ("activityIN") to a threshold to select a filter. For claim 18, Petitioner argued Vehvilainen's threshold (which includes the term "MQUANT") varied depending on the quantizer used in encoding, as MQUANT is a quantization factor. For claim 19, Petitioner contended it would have been obvious to modify Vehvilainen's "less than" comparison to a "greater than" comparison to check if the detail exceeds the threshold, as this is a minor and routine programming change.
    • Expectation of Success: As Vehvilainen taught a complete, working system, a person of ordinary skill in the art (POSITA) would have had a high expectation of success in implementing its teachings to achieve the claimed method.

Ground 2: Obviousness over Vehvilainen in view of Kadono - Claims 1, 2, 4, and 17-19 are obvious over Vehvilainen in view of Kadono.

  • Prior Art Relied Upon: Vehvilainen (’873 patent) and Kadono (Application # 2004/0076237).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that a POSITA would combine Vehvilainen's adaptive deblocking framework with specific techniques taught by Kadono. The combination involved two primary modifications. First, a POSITA would replace Vehvilainen’s "min-max approximation" method for calculating pixel variance with Kadono’s method of calculating the "sum of absolute differences" (SAD) across the block boundary. This directly taught the limitation of claim 2. Second, a POSITA would incorporate Kadono's teaching of using an additional threshold (Ω) to further refine filter selection in smooth areas, allowing for a stronger filter in "super-smooth" regions and a standard filter in "smooth" regions. This modification would improve the adaptiveness of Vehvilainen's system.
    • Motivation to Combine: A POSITA would combine the references to improve the performance and visual quality of Vehvilainen's deblocking method. Both references addressed the same problem in the same technical field. SAD was a well-known, accurate, and computationally efficient method for measuring pixel variation, making it an obvious substitute for Vehvilainen's calculation. Furthermore, Vehvilainen itself suggested that increasing the number of "selection branches" improves adaptiveness, providing an express motivation to add Kadono's additional thresholding step (the Ω threshold) to refine the filtering process for smooth regions.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success. The combination involved substituting or adding well-understood techniques (SAD, thresholding) within compatible systems. Both references were designed for standard block-based video codecs, and the proposed modifications were routine for a POSITA in the field of video processing.

4. Key Claim Construction Positions

  • "determining the level of detail": Petitioner argued this term, found in claim 1, should be understood to encompass determining detail based on the level of variation across pixels. This construction is supported by the specification and would include well-known calculations such as the sum of absolute differences (SAD), which is central to Petitioner's obviousness argument under Ground 2.
  • "pixels on either side of the block boundary": Petitioner asserted this phrase, also in claim 1, is satisfied by selecting pixels on both sides of the boundary. This is critical because the patent owner had previously characterized prior art as only "one dimensional." Petitioner pointed to the patent’s own examples and discussion of the MPEG-4 standard to show that "on either side" was used to mean "on both sides" in the context of the invention.

5. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under §325(d) was inappropriate because the prior art references relied upon (Vehvilainen and Kadono) were not cited or considered during the original prosecution of the ’651 patent.
  • Petitioner also asserted that other factors weighed in favor of institution, noting that this was the first petition filed by Netflix and Hulu against the ’651 patent. While a third party had previously filed a petition, it relied on different prior art, and no institution decision had yet been issued in that proceeding.

6. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 1, 2, 4, and 17-19 of the ’651 patent as unpatentable.