TVision Insights Inc v. Nielsen Co US LLC

1. Case Identification

• Case #: IPR2023-01014
• Patent #: 11,470,243
• Filed: June 16, 2023
• Petitioner(s): TVision Insights, Inc.
• Patent Owner(s): THE NIELSEN COMPANY (US), LLC
• Challenged Claims: 1, 4-6, 8, 9, 11-14, 16, and 18-20

2. Patent Overview

• Title: Methods and Apparatus to Obtain Exposure Data for Media Exposure Environments
• Brief Description: The ’243 patent discloses systems and methods for television audience measurement. The technology combines audio content recognition (e.g., audio signatures) to identify media content with image-based audience analysis (e.g., facial detection and recognition) to identify audience members.

3. Grounds for Unpatentability

Ground 1: Obviousness over Lu-577 and Tian - Claims 1, 4-6, 8, 9, 11-14, 16, and 18-20 are obvious over Lu-577 in view of Tian.

• Prior Art Relied Upon: Lu-577 (Application # 2002/0059577) and Tian (a 2004 conference paper on face resolution for expression analysis).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Lu-577 discloses a complete television audience measurement system that meets most limitations of the independent claims. Lu-577 taught identifying television programs using audio signatures and identifying audience members using "known head location and face recognition software" (by incorporating Patent 4,858,000, or "Lu-000," by reference). Petitioner asserted that Tian, a paper on facial analysis, taught techniques for improving such systems, including determining head pose and performing facial analysis on reduced-resolution images to conserve processing resources. The combination, therefore, taught obtaining a sequence of images, detecting a head, determining its orientation (Tian), and matching the head to a known person to determine audience identification (Lu-577). Dependent claims were allegedly met by Tian's teachings on reducing image resolution for pose estimation and Lu-000's method of generating and comparing facial signatures.
  • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Tian with Lu-577 because Lu-577 expressly invited the use of "known head location and face recognition software." Tian provided well-known, beneficial techniques for improving the performance and efficiency of such software, such as performing certain analyses on sequences of images and using reduced-resolution images to save processing time and resources.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success because the combination involved applying known and tested image processing techniques (Tian) to an established type of system (Lu-577) to achieve predictable benefits in efficiency and accuracy.

Ground 2: Obviousness over Nielsen-372 and Steinberg - Claims 1, 9, and 16 are obvious over Nielsen-372 in view of Steinberg.

• Prior Art Relied Upon: Nielsen-372 (Application # 2010/0274372) and Steinberg (Application # 2010/0066822).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner contended that Nielsen-372 disclosed a local metering system for audience measurement that identified TV programs via audio analysis. However, to identify audience members, Nielsen-372’s "people meter" relied on prompting viewers for manual input. Steinberg was argued to disclose a complete, automated facial analysis system for identifying persons in images, including detecting faces, determining pose/orientation, and performing face recognition against a database of known individuals. The combination allegedly taught a system that captures a sequence of images, detects a head, determines its orientation, and matches it to a known person associated with the environment.
  • Motivation to Combine: A POSITA would combine the references to replace Nielsen-372's intrusive and potentially inaccurate manual-input method for audience identification with Steinberg's more reliable, accurate, and automated facial recognition technology. This substitution would improve data quality without interrupting the viewing experience, directly aligning with the goals of an audience measurement system.
  • Expectation of Success: Success was predictable because it involved the simple substitution of one known audience identification technique (manual input) with another well-known and superior technique (automated facial recognition). Nielsen-372 provided the necessary hardware architecture (e.g., computer and camera) to implement Steinberg's methods.

Ground 3: Obviousness over Nielsen-372, Steinberg, and Tian - Claims 1, 4-6, 8, 9, 11-14, 16, and 18-20 are obvious over Nielsen-372 and Steinberg in view of Tian.

• Prior Art Relied Upon: Nielsen-372 (Application # 2010/0274372), Steinberg (Application # 2010/0066822), and Tian (a 2004 conference paper).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground built upon the combination of Nielsen-372 and Steinberg from Ground 2. Petitioner argued that Tian was added to further improve the resulting system. Tian taught that the accuracy of certain facial analysis tasks, like face detection and head pose estimation, could be maintained even on reduced-resolution images. This incremental teaching addressed limitations in dependent claims related to reducing the resolution of an image to determine head orientation.
  • Motivation to Combine: A POSITA, having already combined Nielsen-372 and Steinberg, would be motivated to incorporate Tian's teachings to make the automated facial analysis more computationally efficient. By performing initial detection and pose estimation on lower-resolution images, the system could save significant processing resources and time, a well-known goal in the field of computer vision.
  • Expectation of Success: A POSITA would have expected success in adding Tian's teachings, as it represented an application of a known optimization technique (processing reduced-resolution images) to an existing facial analysis framework (Steinberg) to gain a predictable improvement in system performance.

4. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under Fintiv was inappropriate because the co-pending district court litigation had been stayed pending the Patent Trial and Appeal Board's institution decision. Petitioner further argued that denial under 35 U.S.C. §325(d) was unwarranted because none of the prior art references relied upon in the petition were cited or considered during the original prosecution of the ’243 patent.

5. Relief Requested

• Petitioner requests institution of inter partes review and cancellation of claims 1, 4-6, 8, 9, 11-14, 16, and 18-20 of the ’243 patent as unpatentable.