Document Image Segmentation for Layout and Font Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional OCR techniques fail to extract document layout and font style information from physical documents, limiting the ability of computer systems to recreate electronic documents that accurately mimic their physical counterparts, as they only generate machine-encoded text without capturing formatting details like location, font size, or text alignment.
Innovation Solution
A method and apparatus that segment document images into components, determine formatting information by comparing features to clusters associated with different document types, and apply OCR to generate machine-encoded text, storing both text and formatting information in data structures to enable accurate reconstruction and manipulation of documents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional OCR techniques are used to convert physical documents to machine-encoded text, then text extraction is achieved, but document layout and font style information are lost
Solution Approach 1:
The document image is divided into multiple segments, where each segment represents a distinct text region. For each segment, the system extracts formatting information including font style, size, and location separately, then associates this information with the corresponding machine-encoded text. This segmentation approach enables comprehensive information extraction without requiring a complete redesign of the OCR system.
Solution Approach 2:
The patent adds a new dimension of formatting information to the traditional OCR output. Instead of only extracting text characters, the system simultaneously extracts and stores formatting attributes (font style, size, location) as additional data dimensions. This transforms the output from simple text strings to enriched text objects with multiple associated properties, enabling document reconstruction.
2Manufacturing precision
If only machine-encoded text is generated from physical documents, then processing speed is maintained, but the ability to recreate electronic documents resembling physical counterparts is limited
Solution Approach 1:
The system performs preliminary extraction of formatting information during the OCR processing stage itself, rather than requiring a separate post-processing step. By identifying and storing font style, size, and location data alongside text recognition, the system prepares all necessary information for accurate document reconstruction in advance, maintaining processing efficiency while improving reconstruction quality.
3Manufacturing precision
If detailed formatting information is extracted and stored for each segment, then document reconstruction quality improves, but data storage requirements increase
Solution Approach 1:
The system extracts only the essential formatting information needed for document reconstruction—specifically font style, size, and location data—rather than storing complete segment images or all possible document attributes. This selective extraction approach captures sufficient detail for accurate recreation while minimizing unnecessary data storage requirements.
Data Source
AI summary
Systems and apparatuses are disclosed for extracting information from document images. An example method includes segmenting a document image into multiple segments and determining formatting information for each segment. Determining formatting information for a segment includes determining one or more features of the segment and comparing the one or more features of the segment to one or more clusters of features associated with different document types. The formatting information for the segment is based on the comparison. The method also includes, for each segment, storing the formatting information in a data structure associated with the segment. The method further includes, for each segment including text to be identified during information extraction, applying OCR to the segment to generate machine-encoded text and storing the machine-encoded text in the associated data structure.