Image Inspection Thresholds from User-Edited Pseudo-Abnormal Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing inspection systems face difficulties in setting a threshold value for abnormality detection as intended by the user, making it challenging to perform accurate inspections on inspection target images.
Innovation Solution
An information processing system that generates a pseudo-abnormal image based on user operations on an original image, sets its abnormality degree as a threshold value, and compares this with the abnormality degree of inspection target images using a learning model to determine normality or abnormality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a fixed threshold value is used for abnormality detection, then the inspection process is simple, but it is difficult to set the threshold value as intended by the user
Solution Approach 1:
The system creates a copy of the normal image and applies user-specified abnormality information to generate a pseudo-abnormal image. This copied and modified image serves as the basis for determining the threshold value, allowing users to define abnormality criteria without complex manual threshold setting.
Solution Approach 2:
The pseudo-abnormal image acts as an intermediary between the user's abnormality specifications and the final threshold value. By processing the normal image through the learning model with added abnormality information, the system mediates the conversion from user intent to actionable threshold without requiring direct threshold manipulation.
2Measurement precision
If manual threshold setting is required, then user control over inspection criteria is high, but the operation becomes complex and difficult
Solution Approach 1:
The system enables users to specify abnormality information in terms of image features (such as adding defects or modifying regions) rather than requiring direct threshold value input. The learning model then automatically processes this information to generate the appropriate threshold, making the system self-adjusting based on user intent.
Solution Approach 2:
Instead of changing threshold parameters directly, users modify image parameters by adding or modifying abnormality information in the pseudo-abnormal image. This parameter transformation approach allows users to control detection criteria through intuitive image-based specifications rather than numerical threshold adjustment.
3Adaptability or versatility
If pseudo-abnormal images are generated based on user operations, then user-intended threshold setting is achieved, but the processing time increases
Solution Approach 1:
The system performs preliminary processing by generating the pseudo-abnormal image and determining the threshold value before actual inspection begins. By pre-computing the threshold based on user-specified abnormality information, the system eliminates the need for repeated threshold calculations during inspection operations.
Solution Approach 2:
The system replaces manual threshold adjustment mechanisms with an automated learning model that processes the pseudo-abnormal image. This substitution of mechanical threshold-setting operations with automated computational processing reduces time loss while maintaining user-defined adaptability.
Data Source
AI summary
An information processing system includes a processor configured to: acquire an original image; receive a user operation on the original image; generate a pseudo-abnormal image containing an abnormal image feature on a basis of the acquired original image and the user operation; set an abnormality degree of the pseudo-abnormal image as a threshold value, the abnormality degree of the pseudo-abnormal image being acquired by inputting the pseudo-abnormal image into a learning model, the learning model being capable of calculating an abnormality degree of an image when the image is input; acquire an inspection target image; and perform an inspection on the inspection target image by comparing an abnormality degree of the inspection target image with the set threshold value, the abnormality degree of the inspection target image being acquired by inputting the inspection target image into the learning model.