PTAB

IPR2017-00357

Samsung Electronics Co Ltd v. Image Processing Technologies LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Image Processing for Target Tracking
  • Brief Description: The ’445 patent describes methods and systems for identifying and tracking a target within a video signal. The technology involves generating histograms based on pixel characteristics (e.g., luminance, speed, position) from video frames to determine and adjust a target's location in real-time.

3. Grounds for Unpatentability

Ground 1: Obviousness over Gilbert in view of Brady - Claims 1, 4, 6, 9, 18, 24, 25, and 27 are obvious over Gilbert in view of Brady (and Altan for claim 18).

  • Prior Art Relied Upon: Gilbert (a 1980 journal article titled A Real-Time Video Tracking System), Brady (Patent 5,761,326), and Altan (a 1992 symposium proceeding).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Gilbert taught a real-time video system for tracking targets like rockets using intensity and projection histograms, meeting many core claim limitations. Brady taught a similar system for tracking vehicles using histograms based on "edgel intensity" to identify target boundaries. Petitioner contended that the combination taught all limitations of independent claims 1 and 24, including receiving a video signal, generating histograms from a first frame to identify a target and its location, and then generating histograms from a subsequent frame to adjust that location. For claim 18, which involves skipping frames, Petitioner added Altan's teaching of intermittent processing (e.g., every sixth frame) to reduce computational load.
    • Motivation to Combine: A person of ordinary skill in the art (POSITA) would combine the similar systems of Gilbert and Brady to improve Gilbert's target identification process. Brady's use of edgel intensity histograms offered a more direct method for locating target edges compared to Gilbert's two-step intensity and projection method. A POSITA would also have been motivated to incorporate Brady’s and Altan's intermittent tracking techniques to reduce the computational burden inherent in real-time image processing, a well-known challenge in the field.
    • Expectation of Success: Success was expected because the combination involved applying known image processing techniques from Brady and Altan to a very similar system (Gilbert) to achieve predictable improvements in efficiency and accuracy.

Ground 2: Obviousness over Hashima in view of Gilbert - Claims 1, 4, 6, 9, 18, 24, 25, and 27 are obvious over Hashima in view of Gilbert (and Altan for claim 18).

  • Prior Art Relied Upon: Hashima (Patent 5,521,843), Gilbert (a 1980 journal article), and Altan (a 1992 symposium proceeding).
  • Core Argument for this Ground:
    • Prior Art Mapping: Hashima taught a system that uses X-Y projection histograms to recognize and track pre-defined, high-contrast target marks to guide a robot arm. Gilbert, in contrast, taught a more generalized system using intensity histograms and probability algorithms to track a wide array of undefined targets. Petitioner asserted their combination met all limitations of the independent claims, with Hashima providing the core histogram-based tracking framework and Gilbert supplying a more advanced and flexible recognition capability.
    • Motivation to Combine: A POSITA would combine the references to overcome the limitations of Hashima's system. Hashima was restricted to pre-defined marks, whereas Gilbert's probability-based algorithm could identify a wide array of targets under various conditions. A POSITA would be motivated to replace or augment Hashima's rigid recognition method with Gilbert's more flexible algorithm, creating a superior system capable of tracking undefined objects without prior marking.
    • Expectation of Success: The combination was expected to be successful as it involved substituting one known target recognition module (from Hashima) with a more advanced but functionally analogous one (from Gilbert) to predictably enhance the system's overall capability and versatility.

Ground 3: Obviousness over Hashima in view of Brady - Claims 1, 4, 6, 9, 18, 24, 25, and 27 are obvious over Hashima in view of Brady.

  • Prior Art Relied Upon: Hashima (Patent 5,521,843) and Brady (Patent 5,761,326).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Hashima's system for tracking pre-marked targets, when improved with Brady's teachings on general object recognition and efficient processing, rendered the challenged claims obvious. Hashima provided the foundational frame-by-frame histogram tracking system, while Brady supplied key inventive improvements that a POSITA would have readily incorporated.
    • Motivation to Combine: A POSITA would combine the references for several reasons. First, they would improve Hashima's specialized system for robotic grasping by incorporating Brady's more general algorithm for recognizing targets based on edgel intensity, removing the restrictive need for pre-marked objects. Second, a POSITA would introduce Brady's intermittent tracking algorithm to Hashima's system to reduce the significant computational burden of its mandatory frame-by-frame processing, a clear path to optimization.
    • Expectation of Success: Success was expected as the proposed modifications involved integrating well-understood techniques (general object recognition, intermittent tracking) from Brady into a similar histogram-based system (Hashima) to solve known problems of limited target scope and high computational load.

4. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1, 4, 6, 9, 18, 24, 25, and 27 of Patent 8,989,445 as unpatentable under 35 U.S.C. §103.