Visual Query Document Retrieval via Character Quality Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search methods require users to input specific terms, making it difficult to locate desired documents accurately, especially when the desired document contains a combination of terms that are rare in other documents, necessitating a system that can process visual queries like images of documents to retrieve the canonical source document.
Innovation Solution
A computer-implemented method using optical character recognition (OCR) on a server system to process visual queries, scoring character quality, identifying high-quality textual strings, and retrieving the corresponding canonical document, which is then sent to the client system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If term-based searching is used, then users can input specific search terms, but it becomes difficult to locate desired documents accurately when the document contains rare term combinations
Solution Approach 1:
The system creates an optical copy (image) of the document and uses this copy as the search query. Instead of requiring users to manually type search terms, the system processes an image of the document directly through OCR and similarity comparison, enabling accurate document retrieval without complex term input
Solution Approach 2:
The system replaces the mechanical process of manual term typing and composition with an automated optical recognition system. The OCR engine and similarity algorithm automatically extract and compare textual content from images, substituting the manual mechanical input process with automated optical and computational processing
2Measurement precision
If visual query processing with OCR is implemented, then document retrieval accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system performs OCR and similarity comparison on only the necessary portions of the visual query to identify matching documents. Rather than processing every possible document in the database with full OCR, the system uses the visual query to generate targeted search terms and compares only relevant documents, reducing overall processing time while maintaining accuracy
Solution Approach 2:
The system pre-processes and indexes documents in advance, creating searchable representations of document content before queries are submitted. This preliminary indexing allows the visual query system to quickly compare against pre-prepared data rather than performing full OCR and analysis on every document during query processing
Data Source
AI summary
A server system receives a visual query from a client system. The visual query is an image containing text such as a picture of a document. At the receiving server or another server, optical character recognition (OCR) is performed on the visual query to produce text recognition data representing textual characters. Each character in a contiguous region of the visual query is individually scored according to its quality. The quality score of a respective character is influenced by the quality scores of neighboring or nearby characters. Using the scores, one or more high quality strings of characters are identified. Each high quality string has a plurality of high quality characters. A canonical document containing the one or more high quality textual strings is retrieved. At least a portion of the canonical document is sent to the client system.