Digital Printing Registration Error Correction via Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital printing processes often result in registration errors within printed images, which existing methods struggle to accurately detect and correct, especially since traditional registration targets are limited to the edge of the substrate and do not account for errors within the image.
Innovation Solution
A method and system that utilize a processor and printing subsystem to select anchor features in a reference digital image, estimate registration errors in printed images using a neural network, and apply corrections to subsequent digital images to be printed, incorporating techniques such as multi-layered convolutional neural networks and deep learning to detect various types of registration errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional registration targets are used at the edge of the substrate, then the device complexity is reduced, but the measurement precision of registration errors within the image deteriorates
Solution Approach 1:
The patent divides the registration correction task into multiple independent color channel processing units. Each color channel (C, M, Y, K) has its own registration error detection and correction mechanism, allowing precise measurement of registration errors within the image while maintaining manageable system complexity through modular architecture
Solution Approach 2:
The patent transitions from traditional 2D edge-based registration targets to 3D multi-layered color channel analysis. By separating and analyzing each color channel independently in the digital domain, the system achieves precise measurement of registration errors throughout the entire image area, not just at edges
2Manufacturing precision
If neural network-based correction is applied, then the manufacturing precision of printed images is improved, but the device complexity and processing time increase
Solution Approach 1:
The patent performs preliminary correction of registration errors in the digital image domain before the actual printing process. By detecting and correcting errors using neural networks on the digital image data, the system prepares corrected image data for printing, improving manufacturing precision without requiring complex real-time adjustments during printing
Solution Approach 2:
The patent replaces traditional mechanical registration adjustment mechanisms with a digital image processing approach. Instead of physically adjusting registration during printing, the system uses neural networks to detect and correct registration errors in the digital domain, substituting mechanical complexity with computational processing
3Manufacturing precision
If registration correction is performed for each color channel independently, then the manufacturing precision is improved, but the processing time increases
Solution Approach 1:
The patent segments the color image into separate color channels (C, M, Y, K) and processes each channel independently for registration correction. This segmentation allows precise alignment of each color channel while parallel processing reduces overall processing time compared to monolithic approaches
Solution Approach 2:
The patent implements continuous correction where registration errors are detected and corrected iteratively for each color channel. The system continuously refines the correction based on detected errors, ensuring high precision while maintaining efficient processing through optimized iterative algorithms
Data Source
AI summary
A method for correcting an error in image printing, the method includes receiving a reference digital image (RDI). Based on a predefined selection criterion, one or more regions in the RDI that are suitable for use as anchor features for sensing the error, are automatically selected. A digital image (DI) acquired from a printed image of the RDI, is received and the one or more regions are automatically identified in the DI. Based on the anchor features of the DI, the error is automatically estimated in the printed image. A correction that, when applied to the DI, compensates for the estimated error, is calculated. The estimated error is corrected in a subsequent digital image (SDI) to be printed, and the SDI having the corrected error, is printed.