Image Inspection System Selective Character Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image inspection systems fail to confirm defective regions on recording media without revealing the contents, especially when defects overlap with characters or are not easily identifiable.
Innovation Solution
An image inspection system and program that selectively extracts character images based on predetermined rules, such as extracting one character with the greatest overlap when defects overlap with characters, or multiple characters adjacent to the defective region when they do not, to determine the defective region without exposing the content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the entire image is displayed to confirm defective regions, then the defective region can be clearly identified, but the content of the image is revealed
Solution Approach 1:
The patent extracts only the defective region from the entire image for display. The extraction unit identifies and isolates the defective area, presenting it separately from the rest of the image content. This allows inspectors to clearly view the defect while the remaining image content remains hidden, resolving the contradiction between defect identification accuracy and content confidentiality.
2Measurement precision
If character images are extracted when defective region overlaps with characters, then the most defective region can be identified, but the content may be partially revealed
Solution Approach 1:
The patent applies local quality by treating the defective region differently from the rest of the image. When a defective region overlaps with characters, only the overlapping portion (the defective part) is extracted and displayed, while the non-defective portions of characters remain hidden. This selective extraction based on local defect characteristics maintains content confidentiality while ensuring accurate defect identification.
3Reliability
If multiple character images are extracted when defective region does not overlap with characters, then sufficient information for defect determination is provided, but the displayed region increases
Solution Approach 1:
The patent extracts multiple character images surrounding the defective region to provide sufficient context for reliable defect determination. By selecting characters adjacent to the defective area, the system ensures that inspectors have enough information to accurately assess the defect without displaying the entire image. This partial extraction approach balances reliability with minimized display area.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There is provided an image inspection system. An image reading unit reals a formed image which is formed on a recording medium based on original image data to generate read image data. A determination unit compares the read image data and the original image data to determine whether or not the formed image includes a defective region. An extraction unit extracts an image of characters present around the defective region from the read image data. An image display device displays the image of the characters and an image of the defective region such that both the images overlap with each other.