Video Scene Classification for Ad Targeting Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current content classification and targeting methods in video content fail to effectively identify and classify individual scenes and objects in real-time, leading to suboptimal placement of advertisements, which can result in lower conversion rates for advertisers.
Innovation Solution
The use of machine learning systems for real-time content classification, combining image recognition, text recognition, and speech recognition to identify and classify scenes and objects within video content, allowing for precise timing of advertisements during relevant content segments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional content classification methods are used, then the system is simpler to implement, but the classification precision and timing accuracy of advertisements deteriorate
Solution Approach 1:
The patent segments video content into discrete scenes and identifies specific objects within each scene (e.g., cars, people, products). This segmentation enables precise classification at the scene level rather than treating the entire video as a single unit, thereby improving advertisement timing accuracy without requiring complete system redesign
Solution Approach 2:
The system performs preliminary classification of video content into scenes and objects before advertisement placement. By pre-identifying relevant content elements and their timestamps, the system prepares classification data in advance, enabling accurate real-time advertisement targeting without adding complexity to the delivery mechanism
2Measurement precision
If broad genre classification is used, then the classification process is faster and simpler, but the targeting accuracy of advertisements deteriorates
Solution Approach 1:
The patent applies different levels of classification granularity to different parts of the video content. Individual scenes and objects receive detailed classification (local quality), while the overall video maintains broad genre categorization. This enables precise advertisement targeting for specific scenes without requiring complete reclassification of the entire video, preserving classification speed
Solution Approach 2:
The system performs classification only on relevant portions of video content (partial action) rather than analyzing every frame uniformly. By focusing computational resources on scene transitions and object appearances, the system achieves high targeting accuracy for advertisement placement while maintaining overall processing efficiency
3Loss of time
If real-time scene and object classification is implemented, then advertisement timing accuracy is improved, but the computational resources and system complexity increase
Solution Approach 1:
The system performs preliminary classification of video content into scenes and objects before advertisement placement. By pre-identifying relevant content elements and their timestamps, the system prepares classification data in advance, enabling accurate real-time advertisement targeting without adding complexity to the delivery mechanism
Solution Approach 2:
The patent introduces an intermediary classification layer that bridges broad genre classification and specific object identification. This intermediary scene-level classification simplifies the computational burden by organizing content hierarchically, enabling real-time processing without requiring direct analysis of every video frame for advertisement timing
Data Source
AI summary
Video content is evaluated to classify one or more scenes or objects of the video content. The classifications may be evaluated against one or more rules for determining whether to include keywords associated with the classifications for targeting supplemental content. Classifications that satisfy the one or more rules may be used for selection of supplemental associated with one or more keywords Selected supplemental content may be included in video content in a break period following primary content.