Image Defect Detection via Region Segmentation and Adaptive Ordering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image defect detection technologies face challenges in performing high-accuracy defect detection within a predetermined time and setting optimal detection conditions for printed materials, leading to inefficiencies and increased costs due to reliance on high-speed processing devices and user-experienced detection criteria.
Innovation Solution
An image defect detection system that divides the original and printed images into regions, calculates image feature amounts, extracts differential strengths, and determines an optimal detection order based on expected defect values, allowing for efficient and accurate defect detection within a predetermined time frame.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If high-speed processing capability is used to end detection process in short time, then detection speed is improved, but device cost increases
Solution Approach 1:
The patent divides the printed material into multiple determination regions and processes each region independently. This segmentation allows the detection system to handle smaller image portions sequentially or in parallel, reducing the computational burden on any single processing unit and enabling the use of standard-speed processors rather than requiring expensive high-speed processing capabilities.
Solution Approach 2:
The patent performs detection on a subset of determination regions rather than requiring complete analysis of the entire printed material. By selectively processing only necessary regions or stopping after a predetermined number of regions are analyzed, the system achieves detection within predetermined time using standard processing speed, avoiding the need for costly high-speed processing equipment.
2Measurement precision
If detection condition is set too severe, then detection accuracy is improved, but number of fail sheets increases and productivity is degraded
Solution Approach 1:
The patent applies different detection conditions to different determination regions based on their characteristics. By adjusting detection sensitivity and criteria locally for each region rather than using uniform severe conditions across the entire printed material, the system maintains high detection accuracy where needed while avoiding unnecessary rejections, thus preserving productivity.
Solution Approach 2:
The patent performs detection on a limited number of determination regions rather than analyzing every region with severe conditions. This partial action approach maintains detection accuracy by thoroughly examining key regions while avoiding the productivity loss that would result from applying severe detection criteria across the entire printed material.
3Productivity
If detection condition is set too loose, then productivity is improved, but detection accuracy is degraded and surplus print is caused
Solution Approach 1:
The patent implements region-specific detection conditions where different determination regions have different detection sensitivities and criteria. This local quality approach ensures that regions requiring high accuracy are thoroughly checked while other regions use more lenient conditions, achieving both high detection accuracy and maintained productivity without causing surplus printing.
4Measurement precision
If test process is extended to ensure high-accuracy examination, then detection accuracy is improved, but detection time exceeds predetermined time
Solution Approach 1:
The patent divides the printed material into multiple determination regions and processes them in sequence or parallel. This segmentation enables the system to achieve high-detection accuracy through thorough examination of individual regions while completing the overall detection process within predetermined time by limiting the scope of analysis to manageable portions.
Solution Approach 2:
The patent performs detection on a subset of determination regions rather than requiring complete analysis of all regions. This partial action approach maintains high detection accuracy by thoroughly examining representative regions while ensuring the detection process completes within predetermined time, avoiding the time loss that would result from exhaustive full-material analysis.
Data Source
AI summary
An image defect detection device that divides an original print image and a print image printed on the basis of the original print image into corresponding regions, acquires an image feature amount of each divided region, extracts a strength of a difference of each divided region between the original print image and the print image, calculates an image defect detection time indicating a time required to detect a defect of each divided region of the print image from the image feature amount and the strength of the difference of each divided region, calculates an expected image defect value indicating a possibility of presence of a defect in each divided region of the print image from the image feature amount and the strength of the difference of each divided region, determines an order of detection of the image defect of the divided region of the print image.