Document Retrieval Apparatus with Dual Thumbnail Display
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document retrieval systems face inefficiencies in retrieving desired documents from large collections, particularly when images are used as retrieval keys, as full-text search methods fail to effectively handle images with extensive graphics and visual data, leading to numerous noise results and requiring user subjectivity for similarity evaluation.
Innovation Solution
A document retrieval technique that includes a document retrieving unit and a thumbnail display unit to efficiently display summaries and contexts of documents matching image retrieval conditions, allowing users to quickly find desired documents by displaying first and second thumbnail images associated with matching pages or image elements in the retrieval result list.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If full-text search is used to retrieve documents, then text-based documents can be effectively searched, but documents containing images and graphics cannot be retrieved
Solution Approach 1:
The patent implements a universal retrieval system that handles both text-based and image-based documents through a single interface. The system extracts features from both text and images, creating a unified search capability that adapts to different document types without requiring separate retrieval mechanisms.
Solution Approach 2:
The patent introduces an image processing unit that acts as an intermediary between the user's image-based query and the document database. This unit extracts visual features from input images and compares them against stored document images, enabling image-to-document retrieval without directly modifying the full-text search infrastructure.
2Adaptability or versatility
If similar image retrieval is performed, then documents containing images can be found, but numerous noise results are returned requiring user subjectivity for evaluation
Solution Approach 1:
The patent segments the retrieval results into multiple categories or groups based on similarity thresholds and document relevance. By dividing the large set of noise results into organized segments, users can more efficiently evaluate results without manually examining every single document, reducing the time required for result evaluation.
Solution Approach 2:
The patent applies different evaluation criteria and presentation styles to different segments of retrieval results. High-similarity results are presented with more detailed information and higher prominence, while lower-similarity results are grouped separately. This local differentiation of quality presentation helps users quickly identify relevant documents without exhaustive manual evaluation.
3Quantity of substance
If a large number of retrieval results are displayed, then more potential matches are shown, but efficiency in finding the desired document decreases
Solution Approach 1:
The patent adds a visual dimension to the retrieval results by displaying thumbnail images alongside document metadata. This transforms the traditional text-only result list into a multi-dimensional display that includes visual content, document properties, and relevance indicators. Users can scan multiple results simultaneously using visual cues, dramatically improving retrieval efficiency without reducing the quantity of displayed results.
Data Source
AI summary
A document retrieving apparatus includes a document retrieving unit adapted to retrieve document data which include images that match an input retrieval condition, a retrieval result list display unit adapted to display, on a display unit, a list display of document data that match the retrieval condition based on the retrieval results of the document retrieving unit, and a thumbnail display unit adapted to display, in the list display by the retrieval result list display unit, a first thumbnail image associated with a page or an image element which matches the retrieval condition, and a second thumbnail image associated with another page or another image element which forms the document data that includes the page or the image element which match the retrieval condition.