Document Reading Device with Geometric Box Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for reading documents with physical deformations, such as wrinkles or bends, are unreliable due to the complexity of digital processing and variable effectiveness, leading to insufficient reliability in accurately interpreting filled checkboxes on distorted documents.
Innovation Solution
A reading device and method that uses image acquisition and digital processing to identify and filter the document, detect zones and boxes using pre-stored masks, and geometrically control the detected boxes to assess reliability, without applying smoothing algorithms, allowing for reliable box detection and filling analysis on deformed documents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If digital smoothing algorithms are used to correct physical deformations, then the document image can be geometrically corrected, but the reading reliability remains insufficient due to variable effectiveness and complex processing
Solution Approach 1:
The patent replaces complex digital smoothing algorithms with a geometric control method based on predefined geometric constraints and relationships between document elements. Instead of using sophisticated image processing to correct deformations, the system detects geometric inconsistencies and directly adjusts the interpretation of box positions and filling states based on expected geometric patterns, thereby achieving reliable reading without complex digital correction.
Solution Approach 2:
The patent pre-establishes geometric models and constraints for document layouts, including expected positions, sizes, and relationships between boxes and zones. By having these geometric rules predefined before document processing, the system can quickly assess whether detected elements conform to expected patterns and make reliable determinations about box filling states without requiring complex real-time correction algorithms.
2Shape
If a flattening device is used to physically flatten the document, then the document can be properly positioned for acquisition, but the device increases cost, requires maintenance, and reading remains difficult if flattening is incorrect
Solution Approach 1:
The patent eliminates the need for mechanical flattening devices by replacing them with a digital geometric control system. The system captures images of documents in their natural state (with wrinkles, bends, and curvatures) and uses geometric constraints and relationships between document elements to reliably determine box filling states without requiring the document to be physically flattened or perfectly positioned.
Solution Approach 2:
The patent introduces geometric models and constraints as an intermediary between the physical document and the reading process. These geometric rules serve as a mediator that allows the system to interpret document content reliably even when the physical document is deformed, bridging the gap between imperfect physical state and accurate digital interpretation without requiring mechanical intervention.
3Reliability
If digital processing is used to correct physical deformations, then the acquired image can be transformed, but the processing is complex and has variable effectiveness depending on the deformations
Solution Approach 1:
The patent segments the document into distinct zones and boxes with predefined geometric relationships. By dividing the document into discrete elements with expected geometric patterns, the system can apply simple geometric constraints to each element and their relationships rather than attempting complex global image correction. This segmentation allows reliable reading by checking whether detected elements conform to expected geometric arrangements.
Solution Approach 2:
The patent changes the approach from correcting image geometry to checking geometric parameters of detected elements against predefined constraints. Instead of transforming the image to correct deformations, the system evaluates geometric parameters (positions, sizes, relationships) of detected boxes and zones against expected patterns, making the reading process reliable without complex processing by focusing on parameter verification rather than geometric correction.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method for reading a document (2), the document (2) corresponding to a document template and comprising printed boxes (8, 14) to be filled in by a user, the method comprising acquiring a digital image of the document (2) and digitally processing the acquired image, the digital processing comprising detecting the boxes (8, 14) on the acquired image by mask matching, each box being detected individually by matching with at least one box mask, the box mask or each box mask representing at least a fraction of a box.