Content Recognition via Text-Media Feature Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current content recognition methods fail to accurately recognize information, resulting in low accuracy due to the inability to effectively associate text and media features.
Innovation Solution
A method and apparatus for content recognition that involves acquiring text and media pieces, performing feature extraction on both, determining feature association measures, adjusting text features based on these measures, and using the adjusted features for recognition, thereby improving the accuracy of content recognition results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional content recognition methods are used, then the process is simple, but the recognition accuracy is low
Solution Approach 1:
The patent merges text features and media features into a unified feature space through feature association measures. By combining multiple feature types (text, image, audio) and adjusting them based on their association degrees, the method creates a more comprehensive representation of content, thereby improving recognition accuracy while managing complexity through systematic integration.
Solution Approach 2:
The patent dynamically adjusts text features based on feature association measures that quantify the relationship between text and media features. This parameter adjustment mechanism allows the system to adaptively weight and modify features according to their relevance, improving recognition precision without requiring complete redesign of the recognition framework.
2Measurement precision
If feature association measures are calculated between all media features and text features, then recognition accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the feature association calculation process by first extracting features separately (text features, media features), then calculating association measures between corresponding feature pairs. This segmentation allows for manageable computation at each stage while achieving comprehensive feature association analysis, balancing accuracy with computational feasibility.
Solution Approach 2:
The patent calculates feature association measures selectively for relevant feature pairs rather than exhaustively for all possible combinations. By focusing computational resources on the most pertinent associations between text and media features, the method achieves sufficient recognition accuracy without unnecessary computational overhead.
Data Source
AI summary
A method for content recognition includes acquiring, from a content for recognition, a text piece and a media piece associated with the text piece, performing a first feature extraction on the text piece to obtain text features, performing a second feature extraction on the media piece associated with the text piece to obtain media features, and determining feature association measures between the media features and the text features. A feature association measure for a first feature in the media features and a second feature in the text features indicating an association degree between the first feature and the second feature. The method further includes adjusting the text features based on the feature association measures to obtain adjusted text features, and performing a recognition based on the adjusted text features to obtain a content recognition result of the content. Apparatus and non-transitory computer-readable storage medium counterpart embodiments are also contemplated.