PTAB

IPR2019-00002

Samsung Electronics Co., Ltd. v. FotoNation Limited

Log In

1. Case Identification

  • Patent #: 8,254,674
  • Filed: October 3, 2018
  • Petitioner(s): Samsung Electronics Co., Ltd.
  • Patent Owner(s): FotoNation Limited
  • Challenged Claims: 29, 35, 36, 40, 41, and 42

2. Patent Overview

  • Title: Method and Apparatus for Red-Eye Detection in an Acquired Digital Image
  • Brief Description: The ’674 patent describes a method and apparatus for in-camera red-eye detection. The system works by locating candidate regions in an image, assessing if their geometry is consistent with an eye, performing further analysis to filter false positives, and then correcting the confirmed red-eye pixels before storing the final image.

3. Grounds for Unpatentability

Ground 1: Claims 29, 35, 41, and 42 are obvious over Wang in view of Shiffer

  • Prior Art Relied Upon: Wang (Patent 6,278,491) and Shiffer (Application # 2006/0257132).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Wang teaches an automatic red-eye detection and reduction system that operates on a digital image. This system includes a red-eye detector that divides a face region into multiple "eye candidate windows" (claimed as "partial face regions") and analyzes them to determine characteristics like the presence of a red pupil. While Wang discloses the software process, it does not explicitly place it in a portable device. Shiffer was argued to cure this deficiency by teaching a portable processor-based device (a digital camera) with an image sensor, lens, and processor programmed to perform red-eye detection using code stored on processor-readable media.
    • Motivation to Combine: A person of ordinary skill in the art (POSITA) would have been motivated to implement Wang's software-based red-eye detection method within the portable digital camera taught by Shiffer. This combination represents the application of a known software process to a known hardware environment to create a commercially desirable product with automated red-eye correction, a predictable result.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success, as combining a known image processing algorithm with a standard digital camera architecture was a well-understood and common practice in the field.

Ground 2: Claims 29, 35, 41, and 42 are obvious over Gaubatz in view of Shiffer

  • Prior Art Relied Upon: Gaubatz (a 2002 IEEE publication) and Shiffer (Application # 2006/0257132).
  • Core Argument for this Ground:
    • Prior Art Mapping: As an alternative primary reference to Wang, Petitioner asserted that Gaubatz teaches a method to automatically detect and correct red-eye in digital images. Gaubatz’s method involves three stages: face detection (using the Viola-Jones system), red-eye detection, and correction. Crucially, Gaubatz narrows the search for red-eye artifacts to a partial face region (the upper half of the detected face) and analyzes it using four metrics: color variation, redness, redness variation, and glint. The combination with Shiffer is identical to Ground 1, with Shiffer providing the portable camera hardware context for implementing Gaubatz's algorithm.
    • Motivation to Combine: A POSITA would have been motivated to incorporate Gaubatz's flexible and efficient red-eye detection algorithm into a portable device like the one described in Shiffer. This would allow a consumer device to perform a "minimally invasive corrective procedure" automatically, which was a known goal in the art.
    • Expectation of Success: The combination was presented as a predictable integration of a known software technique into a standard hardware platform to achieve an improved, more efficient portable camera.

Ground 3: Claims 29, 35, 41, and 42 are obvious over Steinberg in view of Shiffer and Gaubatz

  • Prior Art Relied Upon: Steinberg (Application # 2006/0093213), Shiffer (Application # 2006/0257132), and Gaubatz (a 2002 IEEE publication).

  • Core Argument for this Ground:

    • Prior Art Mapping: Petitioner argued Steinberg teaches an improved red-eye detection system that first determines "image quality characteristics" (e.g., degree of blur) of an acquired image before deciding whether red-eye correction can be applied. However, Steinberg does not teach performing this quality analysis on a partial face region. Gaubatz, as established in Ground 2, teaches analyzing a partial face region (the upper half) to detect red-eye. Petitioner argued a POSITA would combine these teachings to use Gaubatz's partial-region analysis technique to efficiently perform the image quality pre-check taught by Steinberg. Shiffer again provides the portable device framework.
    • Motivation to Combine: A POSITA would combine these references to improve computational efficiency. Analyzing only a partial face region to determine image quality characteristics, as suggested by Gaubatz, would be a faster and less computationally burdensome process than analyzing the entire image, which would be a desirable improvement for the system taught by Steinberg, especially in a resource-constrained portable device.
    • Expectation of Success: The proposed modification was argued to be a predictable use of a known efficiency technique (partial-region analysis from Gaubatz) to improve a known process (image quality pre-filtering from Steinberg).

  • Additional Grounds: Petitioner asserted additional obviousness challenges, primarily by adding a third or fourth reference to the core combinations. These grounds relied on similar rationales. For example, Hong (Patent 6,633,655) was added to teach performing analysis at different resolutions (full and subsampled) to improve efficiency, while Ray (Patent 6,940,545) was added to teach incorporating face recognition to identify specific individuals.

4. Key Claim Construction Positions

  • Petitioner highlighted a potential dispute over the term "to analyze one or more partial face regions" (claim 29). Petitioner asserted the term should be given its plain and ordinary meaning. It preemptively argued against a narrower construction that might be proposed by the Patent Owner, such as one requiring a face detection algorithm specifically trained to detect partially occluded faces. Petitioner noted that its alternative grounds (in a concurrently filed petition) would render the claims obvious even under such a narrower construction.

5. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under §325(d) would be inappropriate. Although the Shiffer reference was considered during prosecution, it was not considered in the new combinations presented in the petition, nor did the examiner have the benefit of the expert declaration of Dr. James Elder.
  • Petitioner also disclosed it was concurrently filing a second IPR petition against the same claims. That petition includes an additional reference (Baluja) to explicitly address a potential, narrower claim construction that the Patent Owner might advance. Petitioner requested the Board institute both petitions to ensure that all meritorious invalidity grounds could be considered, regardless of how the Board ultimately construes the key claim terms.

6. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 29, 35, 36, 40, 41, and 42 of the ’674 patent as unpatentable.