Automated Printed Media Verification via Difference Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current printing evaluation methods for mission-critical applications are inefficient and prone to errors, relying on time-consuming operator verification and spot checks, which can miss quality issues in printed media products, especially in high-stakes industries like healthcare where accurate labeling is crucial.
Innovation Solution
A method and system that compares printed output patterns to stored reference patterns using a computed difference image, evaluating features and defects through an XOR logical operation to determine acceptable proximity, allowing for automatic verification without relying on OCR or simplistic scannability checks, thereby ensuring accurate representation of intended graphic information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If operator verification and spot checks are used, then verification can be performed, but operator time and attention are consumed and errors may be missed
Solution Approach 1:
The patent replaces manual operator verification with an automated image processing system that captures printed media images, computes difference images by comparing against reference images, and automatically evaluates defects. This substitution eliminates operator time consumption while maintaining or improving verification accuracy through consistent automated analysis of all printed media.
Solution Approach 2:
The patent creates digital copies (images) of the printed media and performs verification on these copies rather than requiring operators to physically examine each item. The difference image technique compares the captured image against a reference image copy, enabling rapid automated verification without operator involvement.
2Productivity
If spot checks are performed on portions of print products, then some verification is achieved, but 95% or more of products are overlooked
Solution Approach 1:
The automated image processing system enables continuous verification of all printed media items without interruption or sampling. The system captures, processes, and evaluates images in continuous operation, ensuring every item is examined rather than relying on periodic spot checks that leave most items unverified.
Solution Approach 2:
The system performs verification as part of the printing process itself by capturing images of printed media and immediately computing difference images for defect detection. This preliminary automated verification occurs before products leave the production line, ensuring complete coverage without requiring separate inspection steps.
3Ease of operation
If OCR algorithms are used to estimate correctness, then verification can proceed without reference data, but accuracy is compromised due to inability to detect subtle print defects
Solution Approach 1:
The patent replaces OCR-based text recognition with direct image comparison technology that analyzes the actual visual appearance of printed media. By computing difference images between captured and reference images, the system detects subtle print defects, color variations, and positioning errors that OCR algorithms would miss, maintaining high precision while remaining independent of reference text data.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An evaluation method (10), comprising: receiving (101) a reference pattern corresponding to an output pattern over a network, wherein the output pattern is printed on a medium based on the received reference pattern; rendering (12) a scan based instance of the output pattern, wherein the rendered scan based instance comprises: a set of features at least corresponding to the printed output pattern; and zero or more defect features in addition to the set of features at least corresponding to the printed output pattern, the zero or more defect features comprising zero or more printed defect features; computing (13) a difference image based on a comparison of the rendered scan based instance to the received reference pattern, the computed difference image comprising the zero or more defect features of the rendered scan based instance; and characterized by: evaluating (14) the computed difference image upon the zero or more defect features comprising at least one defect feature in relation to a proximity of at least one feature to a location of one or more pixels of the received reference pattern, wherein the evaluating comprises determining (142), based on the evaluation step (14), that the proximity comprises an unacceptably small separation between a position of one or more pixels of the at least one of the zero or more defect features to the location of the one or more received reference pattern pixels.