Abnormality Detection Using Integrated Spatial Filter Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning algorithms for defect inspection in materials face challenges in determining the degree of abnormality without statistically meaningful threshold values, requiring adjustments of various parameters.
Innovation Solution
An information processing apparatus applies multiple spatial filters to an input image, calculates scores based on model groups with parameters representing target shapes, integrates these scores across filtered images, and determines abnormality using an integrated score.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple spatial filters are applied to determine abnormality, then measurement precision is improved, but device complexity increases due to multiple model groups and parameters
Solution Approach 1:
The patent combines multiple filtered images and their corresponding scores into a single integrated score. This merging process consolidates the information from multiple spatial filters and model groups into one comprehensive abnormality determination metric, improving measurement precision while managing system complexity through integration.
Solution Approach 2:
The integrated score serves as a universal metric that can determine abnormality across different spatial filters and model groups. This multi-functional approach allows a single scoring mechanism to handle various filtering operations and model comparisons, reducing the need for separate determination systems for each filter type.
2Adaptability or versatility
If parameter adjustment is required for abnormality determination, then adaptability is improved, but ease of operation deteriorates due to the need for manual parameter tuning
Solution Approach 1:
The system performs self-adjustment by automatically determining appropriate threshold values for abnormality determination based on the integrated scores. Instead of requiring manual parameter tuning, the system adapts to different inspection scenarios through automated threshold selection, maintaining flexibility while eliminating the operational burden of parameter adjustment.
Solution Approach 2:
The patent automatically changes threshold parameters based on the characteristics of the integrated scores from multiple filters. This dynamic parameter adjustment allows the system to adapt to different inspection conditions without manual intervention, maintaining versatility while improving ease of operation through automation.
Data Source
AI summary
An apparatus and a method are disclosed, each of which applies a plurality of different spatial filters to one input image to generate a plurality of filtered images; calculates, for each of a plurality of pixels included in each of the plurality of filtered image, a score indicating a value determined by a difference from a corresponding one of a plurality of model groups, using the plurality of model groups that respectively correspond to the plurality of filtered images and each including one or more models having a parameter representing a target shape; calculates an integrated score indicating a result of integrating the scores of the respective plurality of pixels corresponding to each other over the plurality of filtered images; and determines an abnormality based on the integrated score.