Smart TV Ad Sequence Identification via Watermark Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audience measurement technologies for television viewing are inaccurate, particularly in measuring advertisements, due to reliance on human logs, automated meters, and Set-Top Box data, which fail to verify screen usage and track advertisements accurately, especially in non-traditional media and dynamic ad insertion scenarios.
Innovation Solution
A system and method utilizing an iSpot Server and Smart TV devices to collect and analyze data on television content, including advertisements, through automated content recognition, to determine view rates and advertisement sequences with high accuracy, using algorithms to identify and categorize advertisements and determine their insertion types, thereby providing reliable audience measurement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If Set-Top Box data is used to measure audience sizes, then data accessibility and installed base are improved, but the ability to verify actual screen usage and content viewing is worsened
Solution Approach 1:
The patent introduces watermarks as an intermediary element embedded in content that acts as a mediator between the content delivery system and the measurement system. These watermarks enable verification of actual content viewing by being detectable by ACR technology in client devices, thus resolving the contradiction between easy data collection and accurate viewing verification.
Solution Approach 2:
The patent replaces the mechanical/manual approach of Set-Top Box logging with an automated optical/acoustic detection system using watermarks and ACR technology. This substitution enables automatic verification of content viewing without relying on device power states, thus improving measurement precision while maintaining ease of data collection.
2Ease of operation
If traditional audience measurement methods are used, then implementation simplicity is improved, but accuracy in measuring advertisements and non-traditional media is worsened
Solution Approach 1:
The patent implements a universal watermarking system that functions across multiple platforms and device types (traditional TV, smartphones, tablets, computers). The same watermark technology and ACR detection mechanism work across all these platforms, enabling accurate advertisement measurement universally without requiring platform-specific solutions, thus maintaining ease of operation while improving precision.
3Device complexity
If program schedules are used to estimate viewed content, then measurement process is simplified, but reliability of content identification is worsened
Solution Approach 1:
The patent replaces the schedule-based estimation system with an active detection system using watermarks and ACR technology. Instead of passively relying on schedule data, the system actively detects and identifies content through watermark recognition in client devices, providing reliable content identification while keeping the measurement process relatively simple through automated detection.
4Device complexity
If Set-Top Box algorithms are used to estimate viewed content, then data collection is simplified, but accuracy in determining actual advertisements viewed is worsened
Solution Approach 1:
The patent introduces watermarks as an intermediary that enables direct detection of advertisement content in client devices. This intermediary system bypasses the need for complex Set-Top Box algorithms and external ad log matching, providing accurate determination of actual advertisements viewed while simplifying the overall data collection process through direct client-side detection.
Data Source
AI summary
Television transmissions are received at multiple locations across a large area. Advertisements may be identified, harvested and tagged from within in the television transmissions. The advertisements may be assigned identifiers; media plans may be determined. A datastructure of observed advertisement sequences may be created; the observed advertisement sequences may be determined to have a probability of being unique above a threshold. The smart TVs may render broadcast content and may report advertisement and/or content identifiers. Reported advertisement sequences may be determined and compared to the datastructure of observed advertisement sequences to identify and/or confirm content rendered by the smart TVs.