PTAB

IPR2017-01228

Samsung Electronics Co Ltd v. Image Processing Technologies LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Image Processing
  • Brief Description: The ’001 patent discloses a process for identifying and tracking a target within a video signal. The system uses one or more histograms derived from image frames to analyze pixel characteristics and track the target on a frame-by-frame basis.

3. Grounds for Unpatentability

Ground 1: Claims 5-13 are obvious over Gerhardt in view of Bassman.

  • Prior Art Relied Upon: Gerhardt (Patent 5,481,622) and Bassman (Patent 6,044,166).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Gerhardt discloses the core process of tracking a target (a user's pupil) on a frame-by-frame basis using pixel intensity histograms. Gerhardt identifies the target by creating "blobs" of pixels and analyzing their properties, including the centroid and the X/Y minima and maxima, which define a "bounding rectangle" around the target (mapping to claim 11). Petitioner further asserted that Gerhardt’s teaching of using a smaller "active window" that can be "incrementally increased" if the pupil is not found teaches the limitation of successively increasing a selected area until the target is found (claim 7). Bassman, which teaches tracking vehicles, was argued to supplement Gerhardt by disclosing a process of tracking a target over time by comparing pixel intensity values across successive frames. Petitioner contended the combination teaches all limitations, including determining a target's center at regular intervals (claims 9-10) and that required system components like a memory, temporal processing unit, and spatial processing unit (claims 5-6, 12-13) are either explicitly taught or inherent in any system that processes video over time.
    • Motivation to Combine: A POSITA would combine the references to improve the reliability and speed of Gerhardt's eye-tracking system. Petitioner argued that Bassman's method of explicitly linking a target's position across frames would make Gerhardt's system more robust by reducing false positives that could arise from issues like a user blinking. This would result in a more reliable and efficient tracking system.
    • Expectation of Success: The combination involved applying known and compatible techniques for target tracking to similar image processing systems, which would have led to predictable and successful results.

Ground 2: Claims 5-13 are obvious over Gilbert in view of Ueno and further in view of Gerhardt.

  • Prior Art Relied Upon: Gilbert (a 1980 journal article, "A Real-Time Video Tracking System"), Ueno (Patent 5,150,432), and Gerhardt (Patent 5,481,622).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner presented Gilbert as the primary reference, describing a real-time system for tracking missiles using pixel intensity histograms and a "tracking window." Ueno was introduced for its teaching of using X- and Y-domain histograms derived from interframe differences to define a target's boundaries using its minima and maxima. Gerhardt was added for its teachings on using an "adaptive thresholding" technique to handle changing light conditions and using a target's centroid for precise location. Petitioner argued this three-way combination discloses all challenged claim elements. For example, the combination teaches forming histograms in multiple domains (intensity from Gilbert; X/Y position from Ueno), determining a target's center (from Gilbert's center-of-area calculation or Gerhardt's centroid), performing this process at regular intervals (inherent in the frame-by-frame operation of all references), and drawing a tracking box (taught by all three references).
    • Motivation to Combine: A POSITA would combine these references to create a more robust, flexible, and efficient version of the system taught in Gilbert. Ueno's simpler method of using X/Y minima and maxima for boundary detection would be a faster and more efficient alternative to Gilbert's more complex center-of-area calculation, especially for less noisy targets. Furthermore, adding Gerhardt's adaptive thresholding technique would make the combined system more robust and capable of tracking targets through changing lighting conditions, improving upon the base Gilbert system.
    • Expectation of Success: A POSITA would have a high expectation of success, as the combination involved integrating complementary and well-understood image processing techniques from the same technical field to achieve predictable improvements in system performance and reliability.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued this petition was not redundant relative to a previously filed IPR (IPR2017-00347) against the same patent. It asserted this filing was necessitated by the Patent Owner adding new infringement contentions for the challenged claims in a parallel district court litigation after the first IPR was filed. Petitioner stated this petition targets different claims and relies on new prior art combinations and arguments not present in the earlier filing, making it non-cumulative.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 5-13 of the ’001 patent as unpatentable.