Video Content Detection Using Object Digest Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in efficiently and accurately detecting unauthorized usage of copyrighted video content on the internet, as existing methods are inefficient and require significant human intervention due to the vast amount of media shared daily, making it difficult to identify modified or copied video content.
Innovation Solution
A computer-implemented method that generates digest information for videos, including objects, their timing, and spatial relationships, which is compared to reference digest information to determine similarity, allowing for automated detection and mitigation of copyright infringement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated object recognition and digest information comparison is implemented, then detection accuracy and efficiency improve, but system complexity increases
Solution Approach 1:
The system segments video content into discrete objects and creates digest information representing key visual elements. This segmentation approach breaks down complex video analysis into manageable components (objects, their properties, temporal relationships), enabling efficient comparison while maintaining detection accuracy. The digest information acts as a compressed representation that captures essential content without requiring full video processing.
Solution Approach 2:
The patent introduces digest information as an intermediary representation between the original video content and the comparison process. This intermediary layer transforms complex video data into standardized object-based descriptors that can be efficiently compared against reference digests, reducing computational complexity while preserving detection capability.
2Measurement precision
If comprehensive digest information including objects, timing, and spatial relationships is generated, then measurement precision improves, but computational requirements increase
Solution Approach 1:
The system extracts only the most relevant visual information from video content to create digest representations. By selecting key objects and their essential properties (identity, timing, spatial relationships) rather than processing all video data, the system achieves high detection precision with reduced computational overhead. The extraction focuses on discriminative features that are sufficient for copyright detection.
Solution Approach 2:
The patent transforms video content into a different parameter space by representing it as structured digest information with specific attributes (object identities, temporal timestamps, spatial coordinates). This parameter transformation enables precise comparison operations while reducing the computational burden of working with raw video pixels and temporal sequences.
3Productivity
If automated detection systems are deployed to monitor vast amounts of online video, then productivity improves, but the ability to detect modified content deteriorates without advanced object recognition
Solution Approach 1:
The system performs preliminary object recognition and digest creation on reference copyrighted videos before deployment. This preliminary action builds a library of expected object sequences and patterns that can be used to detect modifications. By pre-processing reference content into structured digests, the system can efficiently identify alterations in uploaded videos through comparison, maintaining reliability while enabling high-throughput automated detection.
Data Source
AI summary
Techniques detecting usage of copyrighted video content using object recognition are provided. In one example, a computer-implemented method comprises determining, by a system operatively coupled to a processor, digest information for a video, wherein the digest information comprises objects appearing in the video and respective times at which the objects appear in the video. The method further comprises comparing, by the system, the digest information with reference digest information for reference videos, wherein the reference digest information identifies reference objects appearing in the reference videos and respective reference times at which the reference objects appear in the reference videos. The method further comprises determining, by the system, whether the video comprises content included in one or more of the reference videos based on a degree of similarity between the digest information and reference digest information associated with one or more of the reference videos.