Microsoft Corp v. Science Applications Intl Corp

1. Case Identification

• Case #: IPR2019-01360
• Patent #: 9,229,230
• Filed: July 23, 2019
• Petitioner(s): Microsoft Corp.
• Patent Owner(s): Science Applications International Corp.
• Challenged Claims: 1, 4-11, 15, 18-25, 29, and 32-39

2. Patent Overview

• Title: System and Method for Identifying Corresponding Portions of Images
• Brief Description: The ’230 patent discloses a system, typically a heads-up display (HMD) for a soldier, that identifies corresponding portions of images from two independently movable video sources. The system first uses motion data from sensors on each source to perform an initial registration, and then evaluates and refines that registration using the image data itself.

3. Grounds for Unpatentability

Ground 1: Claims 1, 15, and 29 are obvious over Roberts and Dobbie.

• Prior Art Relied Upon: Roberts (a 1997 conference paper on helmet sensor integration) and Dobbie (Patent 6,560,029).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Roberts teaches a system with a helmet-mounted display (HMD), a head orientation sensor (HOS), and a weapon orientation sensor (WOS) that provides motion data to register a weapon's field of view within the soldier's HMD. Dobbie teaches a system that fuses images from two co-located cameras (e.g., thermal and image-intensified) onto an HMD and discloses a controller capable of processing and refining the combined video for parallax. The proposed combination would use a helmet-mounted camera from Dobbie as a first video source and a weapon-mounted camera (implied by Roberts' WOS) as a second, independently movable video source. The controller from Dobbie would receive motion data from Roberts' sensors to perform an initial registration and then use Dobbie's image processing capabilities to evaluate that registration.
  • Motivation to Combine: A person of ordinary skill in the art (POSA) would combine these analogous systems to gain the benefits of both: fusing different types of video imagery (Dobbie) while ensuring it is correctly registered to the user's weapon orientation (Roberts). This would be a predictable combination of known elements to improve a soldier's situational awareness.
  • Expectation of Success: As both references operate in the field of military HMDs and address the problem of displaying relevant visual information, a POSA would have a high expectation of success in combining their respective functionalities.

Ground 2: Claims 1, 4, 15, 18, 29, and 32 are obvious over Roberts and Dobbie in view of Zitová.

• Prior Art Relied Upon: Roberts, Dobbie, and Zitová (a 2003 survey paper on image registration methods).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground builds on the Roberts/Dobbie combination by adding the teachings of Zitová to more explicitly satisfy the "evaluate" limitation of the claims. Zitová teaches that it is "highly desirable" to verify the accuracy of an initial registration by using a second, different registration method, such as one based on image features (control points). The combination would use the motion-based registration from Roberts/Dobbie as a first step and then employ an image-feature-based comparison as taught by Zitová as the second "evaluation" step to refine the alignment. For dependent claims, Zitová's disclosure of "phase correlation" as a "rotationally invariant similarity metric" was argued to meet the limitations of claims 4, 18, and 32.
  • Motivation to Combine: A POSA would be motivated to add Zitová's evaluation techniques to the Roberts/Dobbie system to improve accuracy and reliability. Zitová explicitly recommends using multiple methods to check for alignment errors, providing a clear reason to supplement a motion-based system with an image-based verification step.
  • Expectation of Success: Integrating a well-understood image-processing evaluation technique (Zitová) into an existing image-processing system (Roberts/Dobbie) would be a straightforward task for a POSA.

Ground 3: Claims 1, 4-8, 11, 15, 18-22, 25, 29, 32-36, and 39 are obvious over Roberts and Dobbie in view of Azuma, with or without Zitová.

• Prior Art Relied Upon: Roberts, Dobbie, and Azuma (Application # 2004/0051680), optionally with Zitová.

• Core Argument for this Ground:

  • Prior Art Mapping: This combination adds Azuma's teachings on hybrid tracking to the Roberts/Dobbie system. Azuma discloses a hybrid system that uses image feature tracking (template matching) to correct for drift and other errors inherent in inertial sensor systems (like those in Roberts). This provides a more robust method for both the initial identification and subsequent evaluation steps. Azuma's template matching, which is insensitive to image roll, was asserted to be a "rotationally invariant similarity metric" (claims 4, 18, 32). Azuma's use of operators to find high-contrast features was mapped to the limitations of claims 5, 19, and 33.
  • Motivation to Combine: A POSA would be motivated to incorporate Azuma's hybrid tracking techniques to overcome the known problem of drift in the inertial sensors of the Roberts/Dobbie system. Azuma explicitly teaches that combining inertial and vision-based tracking exploits the complementary nature of the technologies to create a more robust and accurate system.
  • Expectation of Success: Combining these known registration techniques was presented as an obvious design choice to achieve a better-performing system with predictable results.

• Additional Grounds: Petitioner asserted a further ground of obviousness for claims 9-10, 23-24, and 37-38 based on the Roberts/Dobbie/Azuma combination in view of Yang (Application # 2005/0207651). Yang was cited for its teaching of calculating a peak-to-sidelobe ratio (PSR) to assess the quality of a match and for an iterative process of selecting new locations if an initial match fails.

4. Key Claim Construction Positions

• Petitioner argued that the term “a second video source, movable independent of the first video source” should be construed to mean that the two video sources are not in a fixed physical relation to each other and are capable of being moved differently along two or more axes. This construction was central to Petitioner's argument that the combination of a helmet-mounted camera (first source) and a weapon-mounted camera (second source), as taught by the prior art, meets this key limitation of the independent claims.

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under 35 U.S.C. §315(b) would be inappropriate. Although a related litigation involving the ’230 patent was pending (Science Applications International Corp. v. The United States of America), Petitioner argued it was not a "privy" of the defendant, the U.S. Government. Petitioner cited its successful motion to intervene in that case, which was granted on the basis that its interests were not adequately represented by the government. Citing WesternGeco, Petitioner asserted that this lack of adequate representation demonstrates the relationship is not "sufficiently close" to establish privity for IPR estoppel purposes.

6. Relief Requested

• Petitioner requests that the Board institute an inter partes review and cancel claims 1, 4-11, 15, 18-25, 29, and 32-39 of the ’230 patent as unpatentable.