Print Inspection Using Delta-E Machine Learning to Reduce Pseudo-Errors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing optical inspection methods for printed images rely heavily on manual parameterization, leading to pseudo-errors and inefficiencies.

Innovation Solution

A multi-layer architecture for optical inspection using machine learning to detect anomalies, measure defect sizes, and classify defects, incorporating color calibration and customer-specific classification through Delta-E image analysis and neural networks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual parameterization is used in existing optical inspection methods, then the inspection process can be performed with conventional systems, but it leads to pseudo-errors and requires heavy manual intervention

Engineering Contradiction:
Improvedefect detection accuracyVSAvoidmanual parameterization effort
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The system performs self-learning by automatically training machine learning models using captured print images and target images. The processor trains defect detection models and classification models without requiring manual parameterization, enabling the system to autonomously improve its defect detection capabilities while reducing operational complexity

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system transforms the inspection approach by changing from fixed manual parameters to dynamic machine learning-derived parameters. The processor generates defect detection models that automatically determine optimal parameters for comparing print images with target images, eliminating the need for manual parameter setting while improving detection accuracy

Inventive Principle:
Principle #35Parameter changes

2Productivity

If conventional differential image techniques are used, then the inspection can be performed with simple methods, but it leads to pseudo-errors and incorrect defect detection

Engineering Contradiction:
Improveinspection efficiencyVSAvoiddefect detection reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system replaces conventional mechanical differential image techniques with machine learning-based defect detection. The processor uses trained defect detection models to automatically identify defects by comparing print images with target images, substituting simple differential methods with intelligent algorithms that reduce pseudo-errors while maintaining inspection efficiency

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system implements feedback mechanisms where the processor continuously improves defect detection by using captured images to train and retrain models. The classification models receive feedback from defect detection results and automatically adjust parameters to reduce pseudo-errors, creating a self-improving inspection system that maintains high productivity while improving reliability

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3771564B1Print inspection device and method for the optical inspection of a printed image of a printed object
Publication Date: 2025.09.03 BUNDESDRUCKEREI GMBH
  • EP3771564B1 patent drawingFigure 1
  • EP3771564B1 patent drawingFigure 2
  • EP3771564B1 patent drawingFigure 3

AI summary

The invention relates to a print inspection device (100) for the optical inspection of a printed image (120) of a printed object, wherein a target printed image (121) is assigned to the printed object, comprising: a processor (110) configured to determine a plurality of raster cell images (111) for the printed image (120) and the target printed image (121) based on a subdivision of the printed image (120) and the target printed image (121) into raster cells (130); and to determine a Delta-E raster cell image (112) for each raster cell (130) based on a color difference between a raster cell image of the printed image (120) and a raster cell image of the target printed image (121).to determine for each pixel of the Delta-E raster cell image (112) based on a pixel-specific threshold function (140) whether a pixel defect (113) is present, wherein the pixel-specific threshold function (140) is based on a previous training (150) of a database (160) of Delta-E raster cell images with associated manually specified ground-truth cell images; to assemble the previously determined pixel defects (113) in the respective Delta-E raster cell images (112) into a defect image (114) which provides an overview of the pixel defects (113) of the printed image (120); and to output an inspection result (116) based on a user-specific classification (115) of the defect image (114).