OCR Text Grouping via Vertical Alignment Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current OCR software faces difficulties with non-vertically aligned text, leading to erroneous results, especially in documents with mixed fonts, font sizes, and graphical cell borders, resulting in mis-read, double-read, and missing text, which affects the quality and integrity of OCR output.
Innovation Solution
A method that preprocesses images by identifying and grouping text with vertically aligned bounding boxes, allowing for separate OCR processing of aligned text groups, thereby improving OCR efficacy without requiring updates to existing OCR software.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If raster-based OCR scanning is used for text processing, then the OCR process can be simplified and executed efficiently, but erroneous results occur when text is not vertically aligned within rows
Solution Approach 1:
The patent segments the document image into multiple processing zones based on detected text rows and their vertical alignment characteristics. By dividing the image into zones with consistent text alignment properties, the system enables accurate OCR processing for each zone while maintaining overall processing efficiency. This resolves the contradiction by allowing simplified raster scanning within each zone while handling non-aligned text across different zones separately.
Solution Approach 2:
The patent performs preliminary analysis of text row alignment before executing the main OCR process. By pre-identifying vertically aligned text rows and organizing them into zones, the system prepares the image data in advance, allowing the subsequent OCR scanning to proceed efficiently with high accuracy for both aligned and non-aligned text scenarios.
2Reliability
If sophisticated algorithms are implemented in OCR software to accommodate non-vertically aligned text, then OCR accuracy for non-aligned text improves, but software complexity increases and requires new software releases
Solution Approach 1:
The patent implements a preprocessing segmentation approach that divides the document into zones based on text alignment characteristics before OCR processing. This segmentation is performed as a separate image analysis step, allowing the core OCR engine to remain relatively simple while still handling non-aligned text effectively through zone-based processing.
Solution Approach 2:
The patent introduces an intermediary preprocessing step that analyzes text row alignment and creates zone definitions. This intermediary layer acts as a mediator between the input image and the OCR engine, preparing the data in a format that enables accurate recognition without requiring complex modifications to the core OCR algorithm, thus avoiding the need for new software releases.
3Device complexity
If the entire page is scanned and converted at once in full page mode, then the OCR process is simple to implement, but processing time increases and error correction becomes more difficult
Solution Approach 1:
The patent segments the full page into multiple text zones based on vertical alignment characteristics, allowing parallel or sequential processing of smaller zones rather than processing the entire page as a single unit. This reduces processing time while maintaining the relative simplicity of the OCR system through modular zone-based handling.
Solution Approach 2:
The patent performs preliminary zone identification and text row analysis before the main OCR conversion process. By pre-organizing the page content into meaningful zones, the system enables more efficient processing and easier error localization, reducing both processing time and the complexity of error correction while maintaining systematic simplicity.
Data Source
AI summary
A method to improve the efficacy of optical character recognition (OCR) includes scanning an electronically stored representation of a whole or partial document, identifying an image having text in the electronically stored representation of a whole or partial document, identifying the text within the image, and generating a plurality of bounding boxes around the identified text using blob detection. The method also includes grouping together certain text bounding boxes of the plurality of text bounding boxes that are vertically aligned with each other to generate a plurality of aligned text bounding boxes and performing OCR on the aligned text bounding boxes to generate a plurality of OCR groups of text. In addition, the method includes generating a resultant representation of a whole or partial document electronically using the plurality of OCR groups of text and saving the resultant representation of a whole or partial document electronically.