Media Analysis Framework Combining Intelligence Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content intelligence algorithms struggle to combine their results effectively, as they often require post-processing and context determination, making it difficult to use their outputs directly and collaboratively in applications.
Innovation Solution
A software framework that allows for the sequential analysis of digital media assets using raw analyzer modules to generate raw analyzer result data, which is then stored and used by feature algorithm modules to produce enhanced and simplified analysis results, facilitating the combination of content intelligence modules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If content intelligence algorithms are used to analyze digital media assets, then information about the media is provided, but the results require post-processing and context determination making them difficult to use directly
Solution Approach 1:
The patent introduces an intermediary layer (post-processing module and context determination module) between the content intelligence algorithms and the final application. This intermediary automatically combines results from multiple algorithms, determines context, and formats output in a usable form, thereby preserving information completeness while improving ease of use by eliminating manual post-processing requirements
Solution Approach 2:
The patent merges multiple content intelligence algorithm results into a unified output through automated post-processing. By combining results from different algorithms and integrating context determination, the system produces consolidated information that is both comprehensive and directly usable by applications, resolving the contradiction between information completeness and ease of operation
2Reliability
If multiple content intelligence algorithms are combined in a hardwired or orderly manner, then analysis results can be generated, but changes are difficult to implement and the system lacks flexibility
Solution Approach 1:
The patent implements a dynamic architecture where the combination of content intelligence algorithms is not fixed but can be configured and modified based on different analysis needs. The system allows dynamic selection and weighting of algorithms, enabling flexible adaptation to changing requirements while maintaining reliable analysis through proven algorithm combinations
Solution Approach 2:
The patent creates a universal framework that can accommodate multiple content intelligence algorithms and different analysis scenarios through a common post-processing interface. This multi-functional architecture enables the same system to reliably handle various types of media analysis while adapting to different requirements, resolving the contradiction between reliability and flexibility
3Measurement precision
If each content intelligence algorithm provides output in its specific environment, then the algorithm works well in its context, but it is difficult to combine or build upon results from other algorithms
Solution Approach 1:
The patent applies homogeneity by standardizing the output format and interface of multiple content intelligence algorithms through a common post-processing module. Each algorithm maintains its specialized performance in its specific environment, while the standardized interface enables seamless integration and combination of results, reducing integration complexity without sacrificing algorithm performance
Data Source
AI summary
A method for analyzing media assets such as video and audio files. The method includes providing access to all the frames of a digital media asset. The method includes, with a microprocessor, running a raw analyzer modules to analyze the asset frames to produce sets of raw analyzer result data that are stored in a data cache in a file associated with the asset. The sets of raw analyzer results are linked to the raw analyzer modules with unique identifiers. The digital media asset is played for the raw analyzer modules, which concurrently analyze the temporally-related frames. The raw analyzer results are stored as data tracks that include metadata for the asset such as immutable parameters including histograms. The method includes using a feature algorithm module to generate an analysis result, such as face identification, for the digital media asset based on the raw analyzer results accessed by the identifiers.