Search Criterion Determination System for Image Feature Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In image search systems, when multiple images are input as queries, unintended features are often included, leading to irrelevant search results due to the use of all extracted features as search criteria, which can compromise the accuracy of intended search results.
Innovation Solution
A search criterion determination system that extracts features from each image, determines relevant features based on their relationships, and transmits only those features to the search engine, employing a processor to classify and correct features to eliminate contradictions and prioritize intended search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all features extracted from multiple query images are used as search criteria, then the quantity of search criteria increases, but the accuracy of search results deteriorates due to inclusion of unintended features
Solution Approach 1:
The patent extracts and removes unintended features from the set of all extracted features. The feature contradiction elimination unit identifies and eliminates features that contradict user intent, keeping only the relevant features for search criteria. This resolves the contradiction by reducing quantity of search criteria while maintaining or improving accuracy.
Solution Approach 2:
The patent applies different quality standards to different features extracted from query images. Instead of treating all features equally, the system evaluates each feature's relevance to user intent and assigns different weights or inclusion statuses. This allows the system to maintain high-quality (relevant) features while filtering out low-quality (unintended) features.
2Adaptability or versatility
If features from multiple images are combined, then the comprehensiveness of search criteria improves, but the reliability of search results deteriorates due to feature contradictions
Solution Approach 1:
The patent implements a feedback mechanism where the system evaluates the consistency and contradiction of features after extraction. The feature contradiction elimination unit analyzes the relationships between features from multiple images and provides feedback to determine which features should be retained or removed. This feedback loop maintains reliability while preserving comprehensiveness.
Solution Approach 2:
The patent performs preliminary analysis of feature contradictions before finalizing search criteria. By identifying and eliminating contradictory features in advance, the system ensures that only reliable features are used for search, while still maintaining comprehensive coverage of user intent through careful feature selection.
3Loss of time
If all extracted features are used without filtering, then the processing time is reduced, but the productivity of search tasks deteriorates due to irrelevant results
Solution Approach 1:
The patent applies partial filtering of features rather than complete filtering or no filtering. The system processes all extracted features initially but then applies selective elimination of contradictory features. This partial action approach balances processing time with productivity by removing only the harmful contradictory features while retaining most useful features for efficient search.
Data Source
AI summary
A search criterion determination system includes: at least one processor configured to extract, in a case where plural images are inputted as a query, features from elements included in each image, determine at least one feature to be provided to a search engine as an input feature based on a relation between the extracted features, and transmit information based on the input feature to the search engine.