Reference Image Corner Classification for Defect Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing defect inspection methods struggle to accurately align inspection images with reference images due to varying corner rounding amounts applied uniformly across different patterns, leading to misclassification of pseudo defects.
Innovation Solution
A defect inspection apparatus and method that classifies corners of patterns into classes and applies distinct corner rounding processes based on these classes, using an isotropic filter and corner rounding amount maps to align inspection and reference images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a uniform corner rounding amount is applied to all corners of various patterns, then the corner rounding process is simple and fast, but the reference image and inspection image cannot coincide accurately for all patterns
Solution Approach 1:
The patent applies different corner rounding amounts to different corners based on their local characteristics. The corner rounding amount is determined by the distance from the corner to the nearest pattern edge, creating a locally optimized rounding process that improves alignment accuracy for each specific corner while maintaining overall process efficiency.
Solution Approach 2:
The patent changes the corner rounding amount parameter dynamically based on the corner's position and surrounding pattern geometry. By calculating the distance from each corner to the nearest edge and using this distance to determine the rounding amount, the system adapts the rounding parameter to match local pattern characteristics, resolving the contradiction between uniformity and accuracy.
2Productivity
If the corner rounding process is simplified with a uniform rounding amount, then processing time is reduced, but pseudo defects are generated due to misalignment
Solution Approach 1:
The patent performs preliminary calculations to determine the corner rounding amount for each corner before executing the rounding process. By pre-calculating the distance from each corner to the nearest edge and determining the appropriate rounding amount in advance, the system prepares optimized parameters that prevent pseudo defect generation while maintaining efficient processing during the actual rounding operation.
Solution Approach 2:
The system automatically determines the appropriate corner rounding amount for each corner based on its own geometric characteristics without requiring external intervention or complex manual configuration. The corner rounding process uses the pattern's own geometry (distance to nearest edge) to self-determine the optimal rounding parameter, improving both reliability and processing efficiency.
Data Source
AI summary
A defect inspection apparatus includes an imaging mechanism for imaging a sample; an image acquisition circuit configured to generate an inspection image based on image data of the sample imaged by the imaging mechanism; a development circuit configured to generate a developed image from design data; a reference image generation circuit that includes a resizing process circuit configured to perform pattern classification from the developed image, and execute a resizing process for the developed image based on a resizing amount set for each classification, and that is configured to generate a reference image using the developed image after the resizing process; and a comparison circuit configured to compare the inspection image with the reference image.