Image Label Concatenation for Duplicate Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information searching systems, particularly search engines, face challenges in accurately retrieving and labeling digital images due to incomplete and inconsistent labels associated with surrounding text, leading to suboptimal image searching results when dealing with image duplicates.
Innovation Solution
The method involves identifying and analyzing duplicate images using techniques such as histograms, image intensities, or wavelets to concatenate labels from similar images, creating a comprehensive superset of labels that can be assigned to each image, enhancing keyword-based image searching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If labels are assigned to images using surrounding text from documents, then the labeling process is simple and automated, but the labels become incomplete and inconsistent because only a small portion of surrounding text is relevant to the image
Solution Approach 1:
The patent merges labels from multiple duplicate images into a comprehensive superset of labels. By identifying duplicate images and combining their labels, the system achieves complete and consistent labeling while maintaining automation. This resolves the contradiction by merging the simplicity of automated labeling with the completeness of comprehensive labeling.
Solution Approach 2:
The patent uses duplicate images as copies of each other to propagate labels. When one image has a label, its duplicates inherit that label through the copying mechanism, ensuring complete label propagation across all duplicate images without requiring re-labeling of each individual image.
2Adaptability or versatility
If different documents focus on different parts of the image when describing it, then each document provides specific context, but the labels become inconsistent across duplicate images
Solution Approach 1:
The patent merges labels from multiple documents that describe different parts of the same image. By combining labels from all documents associated with duplicate images, the system achieves label consistency while preserving the context specificity from each document. The superset of labels includes all unique labels from all sources.
Solution Approach 2:
The patent creates a universal labeling approach that works across multiple documents and contexts. The superset of labels serves multiple functions: it provides consistent identification, preserves contextual information from different sources, and enables comprehensive search retrieval across diverse document types.
3Device complexity
If image duplicates are not properly handled, then the search system is simpler, but the retrieval accuracy decreases because related images are not associated with all relevant keywords
Solution Approach 1:
The patent merges the processing of duplicate images into a unified approach. By identifying duplicates and applying a single superset of labels to all of them, the system maintains relative simplicity while significantly improving retrieval accuracy. The duplicate handling is integrated into the existing labeling pipeline rather than adding separate complex processing.
Data Source
AI summary
A system analyzes multiple images to identify similar images using histograms, image intensities, edge detectors, or wavelets. The system retrieves labels assigned to the identified similar images and selectively concatenates the extracted labels. The system assigns the concatenated labels to each of the identified similar images and uses the concatenated labels when performing a keyword search of the plurality of images.