PTAB

IPR2017-00355

Samsung Electronics Co Ltd v. Image Processing Technologies LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Process and Apparatus for Tracking a Target in an Input Signal
  • Brief Description: The ’015 patent relates to a process for identifying and tracking a target within a video signal. The system generates one or more histograms based on pixel characteristics (domains) from an image frame to identify the target, draw a tracking box around it, and center the box.

3. Grounds for Unpatentability

Ground 1: Obviousness over Gilbert and Schaming - Claim 6 is obvious over Gilbert in view of Schaming.

  • Prior Art Relied Upon: Gilbert (A Real-Time Video Tracking System, IEEE Transactions, Jan. 1980) and Schaming (Adaptive Gate Multifeature Bayesian Statistical Tracker, 1982).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Gilbert, a 1980 article describing a real-time video tracking system for missiles, disclosed most limitations of claim 6. Gilbert’s system operated on a frame-by-frame basis, using histograms based on pixel intensity (a domain) to separate a target from its background and plume (classes). It then used X- and Y-projection histograms to find the target's center, drew a tracking window, and used a "boresight correction signal" to control telescope optics, thereby centering the tracking window relative to the optical axis. Petitioner contended that Schaming, which describes a statistical tracker, supplied any missing elements. Schaming taught using multiple feature domains simultaneously (e.g., intensity, edge magnitude, spatial frequency) in an N-dimensional histogram, improving upon Gilbert’s primary reliance on intensity. Schaming also explicitly taught adjusting a tracking window and, if the system controls the camera, using an error signal to point the camera so the target remains centered in the frame.
    • Motivation to Combine: Petitioner asserted a POSITA would combine these references because they addressed nearly identical problems in the same field. The authors of the papers were also working in close collaboration on related U.S. Army research projects. A POSITA reading Gilbert’s disclosure of using pixel "features" would have been motivated to consult references like Schaming to implement more advanced features like edge magnitude and spatial frequency. Furthermore, Schaming offered a more generalized and improved statistical method for separating a target from a background, which a POSITA would have found desirable for improving Gilbert's system beyond tracking missiles against a simple background.
    • Expectation of Success: A POSITA would have had a high expectation of success, as the combination involved applying known image processing techniques (multi-domain histograms from Schaming) to a similar, well-understood system (Gilbert) to achieve predictable improvements in tracking robustness and efficiency.

Ground 2: Obviousness over Gilbert and Ueno - Claim 6 is obvious over Gilbert in view of Ueno.

  • Prior Art Relied Upon: Gilbert (A Real-Time Video Tracking System, IEEE Transactions, Jan. 1980) and Ueno (Patent 5,150,432).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground relied on Gilbert for the same core teachings as Ground 1: a frame-by-frame tracking system using histograms to identify a target, draw a tracking window, and center it. Ueno, which described a video encoding system for detecting and tracking human faces, was argued to supplement Gilbert. Ueno explicitly taught creating X- and Y-domain histograms from interframe differences to identify a facial region. It then specified a rectangular tracking box defined by the minima and maxima of these histograms. Petitioner argued that Ueno’s simple rectangular tracking box provided a clear and explicit disclosure for the "drawing a tracking box" limitation.
    • Motivation to Combine: Petitioner argued a POSITA would combine Gilbert and Ueno to improve Gilbert's system. Gilbert’s tracking window was complex, partitioning the view into separate target, plume, and background regions, which could be visually distracting. A POSITA would have been motivated to implement the simpler and more direct tracking box taught by Ueno, which cleanly bounded the target based on its X/Y histogram boundaries. This would reduce visual clutter and computational burden by eliminating the need to process separate plume and background regions. Ueno’s explicit teaching of X- and Y-axis histograms also clarified and provided an efficient implementation for the "projections" described in Gilbert.
    • Expectation of Success: A POSITA would expect success in replacing Gilbert’s complex window with Ueno’s simpler tracking box, as it was a straightforward design choice that would predictably result in a more efficient and user-friendly system.

Ground 3: Obviousness over Hashima and Schaming - Claim 6 is obvious over Hashima in view of Schaming.

  • Prior Art Relied Upon: Hashima (Patent 5,521,843) and Schaming (Adaptive Gate Multifeature Bayesian Statistical Tracker, 1982).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Hashima taught a system for recognizing and tracking a predetermined target mark using a video camera. Hashima’s system processed images frame-by-frame, generated X- and Y-projected histograms, and identified the known target by counting the number of peaks and valleys in the histograms. After detection, Hashima calculated the target’s center from the histogram minima and maxima to guide a robot arm, thereby keeping the camera centered on the target. Petitioner argued that Hashima disclosed all elements of claim 6 except for the use of a plurality of domains beyond X and Y coordinates. Schaming was introduced to supply this teaching, as it explicitly disclosed using multiple features (intensity, edge magnitude, etc.) to create N-dimensional histograms for robust target identification.
    • Motivation to Combine: A POSITA would combine Hashima and Schaming to create a more versatile tracking system. Hashima’s method was limited to identifying known target shapes by matching histogram peak-and-valley patterns. A POSITA would have recognized this limitation and looked to references like Schaming, which taught a more generalized statistical process for identifying targets of unknown or variable shape based on features like intensity and texture. Incorporating Schaming’s algorithm into Hashima’s system would allow it to track a wider variety of targets, representing a significant and logical improvement.
    • Expectation of Success: There would be a reasonable expectation of success in applying Schaming’s generalized statistical detection method to Hashima’s tracking framework, as it involved substituting one target identification module (peak/valley matching) with a more robust and flexible one to achieve the predictable benefit of tracking more diverse targets.

4. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claim 6 of the ’015 patent as unpatentable.