PTAB

IPR2025-01522

Samsung Electronics America Inc v. SnapAid Ltd

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Real Time Assessment of Picture Quality
  • Brief Description: The ’682 patent describes a computerized method for estimating the quality of a digital image. The method involves using data from various sensors (e.g., motion sensors, camera modules) to obtain multiple quality indicators (QIs), applying weights to these indicators, and calculating a total quality value to provide real-time feedback or automated adjustments.

3. Grounds for Unpatentability

Ground 1: Obviousness of Claims 1, 2, 5-8, 11, and 14-16 over Anon, Takeuchi, and Jasinski

  • Prior Art Relied Upon: Anon (Patent 8,508,622), Takeuchi (Application # 2010/0149361), and Jasinski (Application # 2012/0201427).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Anon taught the foundational method of using a "composition and feedback engine" to compute an image quality measure from a weighted combination of parameters, including device motion and camera settings. Takeuchi was alleged to supplement Anon by teaching more advanced object detection, motion analysis (providing a basis for QI3), and a comprehensive point-based system for evaluating image quality. To the extent Anon did not teach dynamically estimating weights, Petitioner asserted Jasinski remedied this by disclosing the dynamic calculation of weights for image regions based on current image data (e.g., scene brightness, image detail), allowing the system to adapt to changing conditions.
    • Motivation to Combine: A POSITA would combine Anon's flexible quality assessment framework with Takeuchi's specific object- and motion-based evaluation to create a more comprehensive and accurate automated system. A POSITA would further incorporate Jasinski's dynamic weighting as a known technique to improve the system's adaptability and accuracy by allowing it to respond to real-time scene conditions, which was a well-understood goal in the field.
    • Expectation of Success: Petitioner contended a POSITA would have an expectation of success because combining these modular, compatible teachings involved applying known techniques to improve a similar system in a predictable manner.

Ground 2: Obviousness of Claims 17, 19, and 20 over Anon, Takeuchi, Jasinski, and Aisaka

  • Prior Art Relied Upon: Anon (Patent 8,508,622), Takeuchi (Application # 2010/0149361), Jasinski (Application # 2012/0201427), and Aisaka (Application # 2010/0246939A1).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground built upon the combination of Anon, Takeuchi, and Jasinski to address the limitations of independent claim 17, which explicitly requires "analyzing the captured image via deep learning algorithms." Petitioner asserted that Aisaka taught using a neural-network-based learning method (i.e., deep learning) to generate a "subject map" by extracting and combining multiple features (luminance, color, motion). This subject map was then used to evaluate the subject's sharpness, thereby providing the claimed analysis via deep learning to obtain a quality value (QI3).
    • Motivation to Combine: A POSITA would be motivated to integrate Aisaka’s deep learning approach into the base system of Anon and Takeuchi to enhance its object detection and image evaluation capabilities. Deep learning was a known and powerful technique for improving the robustness and adaptability of image analysis, making it a logical and desirable improvement for the system taught by the primary references.
    • Expectation of Success: Success would be expected because deep learning was an established technique for image analysis, and the modular frameworks of both Anon and Takeuchi were readily adaptable to incorporate such advanced algorithms for improved performance.

Ground 3: Obviousness of Claim 3 over Anon, Takeuchi, Jasinski, and Alhadef

  • Prior Art Relied Upon: Anon (Patent 8,508,622), Takeuchi (Application # 2010/0149361), Jasinski (Application # 2012/0201427), and Alhadef (Patent 8,009,198).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground targeted dependent claim 3, which requires estimating the first weight (c1, associated with device motion) based on the "precision error, reading resolution, or drift in time" of the motion or location sensor. Petitioner argued that Alhadef taught compensating for inherent sensor inaccuracies, including temporal drift in accelerometers and gyroscopes. Alhadef described using these sensor-specific error characteristics to adjust or weight the contribution of each sensor in determining a device's position, for example through the use of Kalman filters.
    • Motivation to Combine: A POSITA would have been motivated to incorporate Alhadef's teachings to improve the reliability and accuracy of motion estimates in the system established by Anon and Takeuchi. Because the base system relied on sensor data, applying a known technique for sensor error compensation as taught by Alhadef would have been a straightforward and obvious way to improve overall system performance.
    • Expectation of Success: There was a reasonable expectation of success given the well-established practice of sensor fusion and error compensation in the field. Alhadef provided detailed guidance on correcting for sensor errors, and integrating these corrections as weighting factors was a predictable modification.
  • Additional Grounds: Petitioner asserted numerous additional obviousness challenges against various dependent claims by adding other references to the core combination of Anon, Takeuchi, and Jasinski. These included combinations with Ramesh (for probabilistic error estimation), Steinberg (for statistical analysis over time), Liu (for fusing multiple aesthetic algorithms), Cheatle (for using a pre-stored list of faces), Yang (for deep learning), and Ishiwata (for combining gyro and lens module data).

4. Key Technical Contentions (Beyond Claim Construction)

  • Priority Date Challenge: Petitioner argued the ’682 patent was not entitled to its claimed priority dates of October 23, 2012, or February 1, 2013. Petitioner contended that the underlying provisional applications failed to provide written description support for key limitations of the independent claims, such as obtaining a quality value "associated with motion of at least one of said objects" (claim 1) or "analyzing the captured image via deep learning algorithms" (claim 17). This contention, if successful, would establish certain references as prior art under 35 U.S.C. §102.

5. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 1-20 of the ’682 patent as unpatentable under 35 U.S.C. §103.