Masked Image Reconstruction for PCB Defect False-Positive Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing inspection systems using image generation AI for defect detection in printed circuit boards are prone to false positive determinations due to manufacturing variations within acceptable ranges, leading to incorrect identification of defect-free products as defective.
Innovation Solution
A determination apparatus that employs a trained image reconstructing unit to generate multiple reconstruction images by successively overlaying masks on inspection regions, synthesizing these images, and comparing them with the original image to reduce false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If image generation AI is used to detect defects by comparing reconstruction images with inspection images, then new types of defects can be detected, but false positive determinations occur due to manufacturing variations within acceptable ranges
Solution Approach 1:
The mask is divided into multiple regions (first region, second region, third region) with different transparency levels. The first region has higher transparency to preserve manufacturing variations, while the second and third regions have lower transparency to enhance defect detection. This segmentation allows the system to simultaneously maintain reliability for normal variations and improve adaptability for defect detection.
Solution Approach 2:
Different regions of the mask are assigned different transparency qualities based on their functional requirements. The first region (larger area) uses higher transparency to avoid false positives from manufacturing variations, while the second and third regions (smaller areas) use lower transparency to improve defect detection sensitivity. This local differentiation resolves the contradiction between detection capability and determination accuracy.
2Difficulty of detecting and measuring
If a mask is overlaid onto the inspection region to generate a masked inspection image, then defect detection can be performed, but the difference between inspection image and reconstruction image becomes large for normal manufacturing variations
Solution Approach 1:
The mask is segmented into multiple regions with different transparency characteristics. The first region with higher transparency allows normal manufacturing variations to pass through, maintaining measurement precision, while the second and third regions with lower transparency enhance defect detection capability by creating sufficient image differences.
Solution Approach 2:
The transparency parameter of the mask is varied across different regions. By changing the transparency parameter from high in the first region to low in the second and third regions, the system optimizes both defect detection capability and image reconstruction accuracy for different areas of the inspection region.
Data Source
AI summary
To reduce false positive determinations in an inspection system. A determination apparatus includes a trained image reconstructing unit trained such that a first reconstruction image more closely resembles a first image and configured to output the first reconstruction image, in a case where a first mask image is input into the trained image reconstructing unit, the first mask image being generated by overlaying a mask onto an inspection region of the first image, the first image being an image determined not to contain a defect among captured images of an inspection target object, a synthesizing unit configured to generate a second synthesized image by synthesizing a plurality of second reconstruction images in a case where the plurality of second reconstruction images are reconstructed by inputting a plurality of second mask images into the trained image reconstructing unit, the plurality of second mask images generated by a plurality of masks being successively overlaid onto an inspection region of a second image, the second image being a captured image of the inspection target object, and a determination unit configured to compare the second synthesized image with the second image to determine whether or not the second image contains a defect.