PTAB

IPR2017-01189

Samsung Electronics Co Ltd v. Image Processing Technologies LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Image Processing Using Histograms
  • Brief Description: The ’293 patent discloses a device and method for processing image scenes by acquiring and analyzing histograms of pixel parameters. The system uses classifiers and coincidence units to selectively build histograms based on criteria that can be updated in a "learning mode" using statistics derived from the image data.

3. Grounds for Unpatentability

Ground 1: Obviousness over Pirim and Yoda - Claims 3-17 are obvious over Pirim in view of Yoda.

  • Prior Art Relied Upon: Pirim (International Publication No. WO 99/36893) and Yoda (Patent 5,239,594).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Pirim, which discloses a system for detecting driver drowsiness, teaches nearly all elements of independent claim 3. This includes a visual perception processor with multiple histogram calculation units for different pixel parameters (e.g., speed, direction), a data bus, a time coincidences bus, classifiers for comparing pixel data to a criterion, and a test unit for calculating histogram statistics (MIN, MAX, etc.).
    • Motivation to Combine: Petitioner contended that while Pirim discloses that its classification criteria can be modified by a computer, it does not specify the process. Yoda, directed to a similar image analysis system, explicitly teaches a "learning mode" where classification criteria (weighting vectors) are updated to improve performance. A Person of Ordinary Skill in the Art (POSITA) would combine Yoda’s well-defined learning mode with Pirim's system to improve its adaptive capabilities and classifier performance, as it was a known technique for enhancing such systems.
    • Expectation of Success: A POSITA would have a reasonable expectation of success in applying Yoda's learning mode concept to Pirim’s system, as both operate on similar principles of image classification, and the integration would involve applying a known improvement technique to a similar system.

Ground 2: Obviousness over Pirim and Eriksson - Claims 20-21 are obvious over Pirim in view of Eriksson.

  • Prior Art Relied Upon: Pirim (International Publication No. WO 99/36893) and Eriksson (a 1998 IEEE publication titled “Eye-Tracking for Detection of Drive Fatigue”).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Pirim discloses the core device recited in the base claims 18 and 19, including a controller, data processing block, histogram unit, and classification unit for detecting driver drowsiness. The arguments rely on Pirim's disclosure of a controller that can read histogram statistics and control classification registers.
    • Motivation to Combine: Both Pirim and Eriksson are directed to the same technical field: using histograms to track a driver's eyes to detect fatigue. Petitioner argued that Eriksson solves the specific problem of setting appropriate classification thresholds by teaching an "adaptive thresholding" algorithm. This algorithm automatically adjusts thresholds based on whether "good eye-regions" are found. A POSITA would be motivated to apply Eriksson's adaptive thresholding to Pirim's system to improve its robustness under varying lighting conditions, a problem explicitly mentioned by Pirim.
    • Expectation of Success: Combining Eriksson's adaptive technique with Pirim's comprehensive system would have been a straightforward application of a known solution to a known problem within the same field, leading to a high expectation of success.

Ground 3: Obviousness over Pirim - Claims 2, 23, and 28 are obvious over Pirim.

  • Prior Art Relied Upon: Pirim (International Publication No. WO 99/36893).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Pirim, standing alone, renders these claims obvious. Pirim was shown to disclose a visual perception processor for automatically detecting events over time using a digital signal (claim 2), a device for detecting visual phenomena (claim 23), and a method of analyzing parameters (claim 28). Pirim’s disclosure of an "anticipation function" that uses statistical information (speed and direction) from successive frames to automatically modify a bounded search area was argued to meet the limitations related to anticipating future parameter values.
  • Additional Grounds: Petitioner asserted an additional obviousness challenge for claims 24-27 over the combination of Pirim and Qian (International Publication No. WO 99/35606). This ground argued for implementing Pirim’s anticipation function using Qian’s specific method of calculating object velocity based on the change in a histogram’s median value between frames.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued that this petition is not cumulative of a previously filed petition (IPR2017-00336) against the same patent. This second petition was necessary because the Patent Owner amended its infringement contentions in parallel district court litigation to assert additional claims not challenged in the first petition. Therefore, discretionary denial was argued to be inappropriate as the petition was filed to challenge all remaining claims that the Patent Owner was likely to assert.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 2-17, 20-21, and 23-28 of Patent 6,959,293 as unpatentable.