Periodic Activity Pattern Detection in Audio Video Content
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Targeted Substitutional Advertising systems fail to effectively detect periodic activity patterns in audio video content viewing habits, particularly missing significant insights from occasional or short-duration viewing by young children, which can be valuable for targeted advertising.
Innovation Solution
A method involving recording activity data in an activity log, suppressing non-relevant data, applying signal processing techniques like Discrete Fourier Transform to convert data into frequency responses, and analyzing these responses to detect periodic patterns, which can include weekly viewing patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If activity data from all viewing sessions is recorded and analyzed equally, then dominant viewing profiles are accurately captured, but periodic patterns from occasional viewers (e.g., children watching weekend programming) are lost
Solution Approach 1:
The patent replaces traditional demographic profiling methods with signal processing techniques. Activity log data is transformed from time-domain representations into frequency-domain representations using Fourier transforms, enabling periodic patterns to be detected through spectral analysis rather than conventional statistical methods
Solution Approach 2:
The patent changes the representation parameters of viewing data by applying frequency transformations. By converting activity logs into frequency responses, the system can identify periodic patterns through spectral peaks, fundamentally changing how viewing behavior is characterized and analyzed
2Reliability
If only large amounts of viewing data are considered for profile derivation, then dominant demographic profiles are stable, but periodic viewing by occasional members (e.g., children) is ignored
Solution Approach 1:
The patent substitutes conventional threshold-based filtering with frequency-domain analysis. Instead of discarding small viewing amounts as noise, the system transforms data into frequency responses where periodic patterns emerge as distinct spectral peaks, allowing reliable detection of recurring viewing behavior regardless of total viewing duration
Solution Approach 2:
The patent introduces frequency response analysis as an intermediary between raw activity logs and demographic profiles. This intermediary transformation layer converts time-domain viewing data into frequency-domain representations, enabling the detection of periodic patterns that would otherwise be obscured in aggregate viewing statistics
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for more accurate detection of periodic viewing patterns, enabling targeted advertising, scheduling of content, and optimization of electronic program guides, thereby improving advertising relevance and viewer engagement.
Implementation Method 1
passing the one or more sets of suppressed activity data through a signal processing function to convert the one or more sets of suppressed activity data to one or more frequency responses
Data Source
AI summary
A method of detecting periodic activity patterns associated with the viewing of audio video content is described. The method includes: recording activity data in an activity log; suppressing the activity log one or more times to suppress non-relevant activity data thereby producing one or more sets of suppressed activity data; passing the one or more sets of suppressed activity data through a signal processing function to convert the one or more sets of suppressed activity data to one or more frequency responses; and analyzing the one or more frequency responses to detect the periodic activity patterns. Related apparatus and methods are also described.