Visual Query Document Retrieval via OCR and Structural Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional text-based searching requires users to explicitly input search terms, making it cumbersome to locate specific documents, especially when the desired document contains a combination of terms found in few other documents, necessitating long query strings without errors.
Innovation Solution
A computer-implemented method processes visual queries by performing optical character recognition (OCR) to extract textual characters and structural information, scoring them to identify high-quality textual strings, and retrieving a canonical document consistent with this information, allowing for the retrieval of a canonical source document from a visual query.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text-based searching is used, then users can input search terms explicitly, but it becomes cumbersome to locate specific documents requiring long query strings without errors
Solution Approach 1:
The patent uses OCR technology to create a visual copy of the document's text and layout. Users can capture an image of the target document, and the system performs OCR to extract and match the visual text representation against database documents, eliminating the need for manual text input while maintaining high retrieval accuracy
Solution Approach 2:
The patent introduces structural information (such as text positioning, formatting, and layout characteristics) as an intermediary between the visual query and the text database. This intermediary enables the system to match documents based on both content and visual structure, improving accuracy without requiring users to manually encode complex search criteria
2Measurement precision
If OCR is performed on visual query to extract text, then document retrieval accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The patent segments the visual query processing into distinct stages: initial OCR extraction of text, separate extraction of structural information (layout, positioning, formatting), and subsequent matching against database documents. This segmentation allows parallel processing of text and structural features, reducing overall processing time while maintaining high accuracy
Solution Approach 2:
The patent performs preliminary OCR processing and structural information extraction on the visual query before the actual document matching begins. By preparing these features in advance, the system reduces the computational burden during the matching phase, thereby decreasing total query processing time
Data Source
AI summary
A server system receives a visual query from a client system, performs optical character recognition (OCR) on the visual query to produce text recognition data representing textual characters, including a plurality of textual characters in a contiguous region of the visual query. The server system also produces structural information associated with the textual characters in the visual query. Textual characters in the plurality of textual characters are scored. The method further includes identifying, in accordance with the scoring, one or more high quality textual strings, each comprising a plurality of high quality textual characters from among the plurality of textual characters in the contiguous region of the visual query. A canonical document that includes the one or more high quality textual strings and that is consistent with the structural information is retrieved. At least a portion of the canonical document is sent to the client system.