Image Re-ranking via Anchor Concept Graph Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image retrieval methods, including keyword-based and content-based image retrieval, often return unsatisfactory results due to user errors and the 'semantic gap' in visual information, leading to a need for re-ranking search results to better align with user intentions.
Innovation Solution
An image re-ranking method that involves calculating an anchor concept set, forming an anchor concept graph, training a classifier, and using concept projection vectors to rank images based on ACG distance, thereby improving the accuracy of search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If keyword-based image retrieval is used, then image semantic matching accuracy is improved, but search result relevance deteriorates due to user input errors and multiple matching contents
Solution Approach 1:
The patent implements feedback mechanisms by analyzing user interaction data with search results and using this information to refine and adjust search algorithms. The system continuously learns from user behavior patterns to improve matching accuracy and filter irrelevant results, thereby resolving the contradiction between precise semantic matching and overall result relevance.
Solution Approach 2:
The patent dynamically adjusts search parameters such as keyword weighting, matching thresholds, and result filtering criteria based on user input quality assessment. By changing these parameters adaptively, the system maintains high semantic matching accuracy while improving the reliability of search results by suppressing irrelevant matches caused by user errors.
2Adaptability or versatility
If content based image retrieval (CBIR) is used, then visual information utilization is improved, but retrieval effect deteriorates due to visual information diversity and semantic gap
Solution Approach 1:
The patent introduces intermediate processing layers including feature extraction modules, similarity computation algorithms, and result ranking mechanisms that bridge the semantic gap between visual information and user intent. These intermediary components translate diverse visual features into standardized representations, enabling reliable retrieval while maintaining adaptability to different visual content types.
3Speed
If traditional image search methods are used, then search speed is improved, but user effort increases due to unsatisfactory search results requiring more viewing
Solution Approach 1:
The patent performs preliminary actions by pre-processing and indexing image data with multiple feature descriptors, pre-computing similarity metrics, and organizing results in ranked order before user interaction. This preliminary preparation enables fast search response while reducing user effort, as results are already optimized and filtered according to relevance criteria.
Data Source
AI summary
The present disclosure relates to an image re-ranking method, which includes: performing image searching by using an initial keyword, obtaining, by calculation, an anchor concept set of a search result according to the search result corresponding to the initial keyword, obtaining, by calculation, a weight of a correlation between anchor concepts in the anchor concept set, and forming an anchor concept graph ACG by using the anchor concepts in the anchor concept set as vertexes and the weight of the correlation between anchor concepts as a weight of a side between the vertexes; acquiring a positive training sample by using the anchor concepts, and training a classifier by using the positive training sample; obtaining a concept projection vector by using the ACG and the classifier; calculating an ACG distance between images in the search result corresponding to the initial keyword; and ranking the images according to the ACG distance.