Binarization Device for Payment Documents Using Window Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current binarization devices are inadequate for accurately processing payment or accounting documents with sensitive data, particularly when printed by dot matrix printers, as they often confuse sensitive data with the background, leading to data loss and inefficiencies in processing and storage.
Innovation Solution
A device and method that utilize primary binarization, memory for form identification, window identification, background estimation, and subtraction to isolate and filter sensitive data pixels, employing morphological operators for accurate binarization and merging of images to enhance data extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If binarization is performed on scanned documents with medium resolution (200-300 dpi), then processing time is reduced and information transmission is limited, but sensitive data may be lost when print quality is poor
Solution Approach 1:
The binarization process is segmented into multiple stages: first binarization of the entire document, then identification of data windows containing sensitive data, followed by secondary binarization specifically for those windows. This segmentation allows different processing strategies to be applied to different regions, improving overall reliability without sacrificing productivity.
Solution Approach 2:
The system performs preliminary binarization and data window identification before final sensitive data extraction. By pre-identifying which regions contain sensitive data and preparing binary masks in advance, the system optimizes the main processing step and ensures no data is lost during the critical binarization phase.
2Measurement precision
If images are captured at high resolution and color, then binarization accuracy improves, but processing time becomes too long and image files become large
Solution Approach 1:
Instead of applying high-resolution processing to the entire document, the system applies enhanced processing only to specific data windows where sensitive data is located. The majority of the document is processed at standard resolution, while only critical regions receive the computationally intensive secondary binarization treatment.
Solution Approach 2:
The system extracts and isolates sensitive data regions from the rest of the document through data window identification. By separating the sensitive data extraction process from general document processing, the system can apply specialized high-accuracy binarization only where needed, rather than processing the entire high-resolution image.
3Ease of manufacture
If known binarization devices are used on documents printed by dot matrix printers, then standard processing is maintained, but sensitive data is confused with background and data is lost
Solution Approach 1:
The system introduces an intermediary secondary binarization process specifically for data windows, which acts as a mediator between the standard binarization and the final sensitive data extraction. This intermediate step uses morphological operations and adaptive thresholding to recover sensitive data that was lost in the first binarization, while maintaining compatibility with standard processing pipelines.
Solution Approach 2:
The system changes binarization parameters specifically for data window processing, using adaptive thresholding and morphological operations tailored to the characteristics of sensitive data prints. These parameter adjustments allow the system to handle poor quality dot matrix prints effectively without compromising the simplicity of standard document processing.
Data Source
AI summary
A binarization device for payment or accounting documents including sensitive data located in respective data window provides a primary binarized document file of the document; a memory stores identification files including identifying images and location information associated to given types of documents; the data window can be identified and localized as comparison with the identification files of the memory; the contribution of the background is subtracted from the window file, the window file is binarized and filtered for spurious pixels obtaining a binarized window file; the binarized document file and the binarized window file are merged to provide the binarized window file in the data window; the evidence of the significant pixel is obtained by sequential analysis on groups of pixels, applying morphological expansion operators on each group of pixels and following erosion of said group of pixels.