PTAB

IPR2017-00353

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 ’134 patent relates to a process for identifying and tracking a target within an input signal, such as a video feed, by generating and analyzing one or more histograms derived from pixel characteristics within each image frame.

3. Grounds for Unpatentability

Ground 1: Claims 1-2 are obvious over Gilbert in view of Hashima.

  • Prior Art Relied Upon: Gilbert (A Real-Time Video Tracking System, IEEE Transactions on Pattern Analysis and Machine Intelligence (Jan. 1980)) and Hashima (Patent 5,521,843).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Gilbert, a 1980 article describing a missile and aircraft tracking system, discloses the core process of tracking a target in a video signal on a frame-by-frame basis using histograms. Gilbert’s system uses histograms based on pixel intensity (gray-scale levels) to separate a target from its background and calculate a center point. However, Petitioner contended that Gilbert’s method for finding the target’s center is computationally complex. Hashima was argued to supply the missing element of claim 1[c] by explicitly teaching the use of X- and Y-projected histograms to determine a target’s boundaries (X and Y minima and maxima) and using that data to calculate a precise center point with a simple formula. For claim 2, Petitioner asserted that Gilbert’s disclosure of a “tracking window” placed around the target image meets the limitation of drawing a tracking box.
    • Motivation to Combine: A POSITA would combine Hashima with Gilbert to improve Gilbert’s system. Specifically, a POSITA would replace Gilbert’s computationally intensive center-finding method with Hashima’s more efficient technique of using X/Y minima and maxima. This would increase processing speed and reduce computational load. Further, because Gilbert suggests that features other than intensity could be used to form histograms, a POSITA would be motivated to incorporate Hashima’s use of X- and Y-domain histograms to improve tracking robustness.
    • Expectation of Success: Petitioner asserted a POSITA would have a reasonable expectation of success, as the combination involved applying known, predictable techniques for histogram analysis and target-center calculation to similar image processing systems.

Ground 2: Claims 1-2 are obvious over Hashima in view of Ueno.

  • Prior Art Relied Upon: Hashima (Patent 5,521,843) and Ueno (Patent 5,150,432).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Hashima discloses the primary elements of claim 1, including tracking a target using X- and Y-projected histograms to find its center based on its boundaries (minima and maxima). Ueno, which teaches facial recognition in a video signal, was argued to provide a more explicit disclosure for certain limitations and particularly for claim 2. Ueno teaches creating X- and Y-histograms from interframe differences and using the minima and maxima of those histograms to define and draw a rectangular tracking box around the detected facial region on a monitor.
    • Motivation to Combine: A POSITA would combine Ueno with Hashima to enhance the system’s user interface and functionality. While Hashima’s system tracks a target, it lacks a clear method for visually highlighting the tracked target for a user. A POSITA would be motivated to incorporate Ueno’s teaching of drawing a tracking box to provide a user with clear visual confirmation that the correct object is being tracked, thereby making the system more user-friendly and effective, especially in visually noisy environments.
    • Expectation of Success: Petitioner argued that success would be predictable, as implementing a tracking box display is a straightforward addition to an image processing system that already identifies the target's boundaries.

Ground 3: Claims 1-2 are obvious over Gilbert in view of Ueno.

  • Prior Art Relied Upon: Gilbert (A Real-Time Video Tracking System, IEEE Transactions on Pattern Analysis and Machine Intelligence (Jan. 1980)) and Ueno (Patent 5,150,432).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Gilbert teaches the foundational tracking process using histograms. Ueno was asserted to supplement Gilbert by teaching the specific method of determining X/Y minima and maxima from histograms to define target boundaries, a technique only suggested as a possibility in Gilbert. Furthermore, Ueno explicitly teaches drawing a simple rectangular tracking box around the target, which Petitioner argued is a direct improvement over Gilbert’s more complex and visually distracting “tracking window” that partitions the frame into separate target, background, and plume regions.
    • Motivation to Combine: A POSITA would be motivated to modify Gilbert’s system with Ueno’s techniques for two primary reasons. First, to implement a simpler and more computationally efficient method of defining the target’s boundaries using X/Y minima and maxima. Second, to replace Gilbert’s visually distracting, multi-region tracking window with Ueno’s simple rectangular box, which would minimize visual distraction for the user and eliminate unnecessary processing steps required to define the background and plume areas.
    • Expectation of Success: Petitioner contended a POSITA would expect a predictable improvement in both computational efficiency and user experience by integrating Ueno’s more streamlined methods into Gilbert’s foundational tracking framework.

4. Relief Requested

  • Petitioner requests institution of an IPR and cancellation of claims 1-2 of the ’134 patent as unpatentable.