Calibration Image Comparison for Quantization Error Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image comparison methods in printing operations are prone to erroneous detections due to quantization errors and long processing times, especially when comparing pre-calibration and post-calibration images, which can result in incorrect identification of image differences.
Innovation Solution
An image comparison method that generates a difference image, applies contraction and outline processing, identifies edge regions, and uses threshold-based pixel comparisons to accurately detect differences between multi-valued images, reducing erroneous detections and processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image comparison methods are used to compare pre-calibration and post-calibration images, then differences can be detected, but erroneous detections occur due to quantization errors and jaggies from RIP processing
Solution Approach 1:
The patent segments the image comparison process into multiple processing stages: generating difference images, creating contracted images through downsampling, generating outline images, and performing edge region analysis. This segmentation allows each stage to handle specific aspects of the comparison, reducing the impact of quantization errors at any single stage and improving overall detection reliability
Solution Approach 2:
The patent introduces contracted images as an intermediary representation between the original pre-calibration and post-calibration images. By downsampling the images to create contracted versions, the system creates a buffer that reduces the impact of high-frequency quantization errors and jaggies while preserving the essential structural information needed for accurate difference detection
2Measurement precision
If detailed pixel-by-pixel comparison is performed to ensure accurate detection of all differences, then detection precision improves, but processing time increases significantly
Solution Approach 1:
The patent divides the image comparison task into sequential segments: first generating difference images to identify potential difference regions, then creating contracted images for efficient overview analysis, followed by outline image generation to define difference boundaries, and finally edge region analysis for precise characterization. This segmentation enables the system to process images efficiently by focusing computational resources on relevant regions rather than performing exhaustive pixel-by-pixel comparison across the entire image
Solution Approach 2:
The patent performs partial comparison by first analyzing contracted (downsampled) images to identify difference regions, then focusing detailed analysis only on those specific regions in the original images. This partial action approach avoids the excessive computational cost of analyzing every pixel in the full-resolution images while still maintaining high detection precision for actual differences
3Quantity of substance
If conventional comparison methods detect all pixel differences, then comprehensive difference detection is achieved, but actual corrections cannot be distinguished from erroneous detections
Solution Approach 1:
The patent extracts and analyzes edge regions specifically from the difference images. By identifying and focusing on edge regions—areas where actual corrections are most likely to occur—the system separates meaningful difference information from spurious differences caused by quantization errors. This extraction approach allows the system to distinguish actual corrections from erroneous detections by analyzing the characteristics of edge regions rather than treating all pixel differences equally
Data Source
AI summary
After a difference image between original image and calibration image is generated, a candidate image of an edge region is generated. Thereafter, an edge image is generated based on the original image, the calibration image, and the candidate image. At that time, the processing target pixel (pixel constituting the candidate image) is determined to be a pixel constituting the edge region when a difference between a pixel value of at least one of nine comparison target pixels in the original image and a pixel value of the processing target pixel in the calibration image is less than or equal to a first threshold value, and when a difference between a pixel value of at least one of nine comparison target pixels in the calibration image and a pixel value of the processing target pixel in the original image is less than or equal to the first threshold value.