Charged Particle Beam Defect Detection via Feature Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing defect determination methods in charged particle beam imaging are oversimplified, leading to frequent misjudgments during sample inspection, as they only consider feature gray values and fail to account for multiple factors.
Innovation Solution
The method transforms charged particle microscopic images into multiple feature images using image transformation operators, allowing for comparison of these feature images to determine defects, thereby reducing misjudgment by evaluating distances between target and reference images using distance evaluation operators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only global feature gray values are used for defect determination, then the inspection process is simple and fast, but the accuracy of defect detection is low and misjudgment occurs frequently
Solution Approach 1:
The patent divides the image analysis process into multiple segments by extracting different feature types (edge features, texture features, shape features) from the image. Each feature type is processed and compared separately, allowing the system to capture various aspects of potential defects rather than relying on a single global gray value metric.
Solution Approach 2:
The patent transitions from one-dimensional gray value comparison to multi-dimensional feature space comparison by incorporating multiple feature types (edge, texture, shape). This dimensional expansion enables more comprehensive defect characterization and reduces misjudgment by considering multiple attributes simultaneously.
2Measurement precision
If multiple feature factors are considered for precise comparison, then defect determination accuracy improves, but the computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary feature extraction and organization before the actual comparison process. By pre-processing the image to extract edge features, texture features, and shape features in advance, the system prepares the data in a structured format that facilitates faster comparison and reduces computational burden during the defect determination phase.
Solution Approach 2:
The patent segments the comparison process into independent feature-type comparisons (edge feature comparison, texture feature comparison, shape feature comparison). This segmentation allows parallel processing of different feature types and avoids the computational overhead of analyzing all features simultaneously, thereby reducing overall processing time.
3Reliability
If simple gray value comparison is used, then the inspection system is easy to operate, but misjudgment occurs frequently due to oversimplification
Solution Approach 1:
The patent implements automated feature extraction and comparison algorithms that perform the complex analysis tasks without requiring manual intervention. The system automatically extracts multiple feature types, compares them against reference images, and determines defects based on the综合分析 of all features, maintaining ease of operation while significantly improving reliability.
Solution Approach 2:
The patent incorporates a feedback mechanism where the comparison results from multiple feature types are integrated to form a comprehensive defect determination. The system evaluates edge features, texture features, and shape features together, using the combined information to reduce misjudgment and improve the reliability of defect detection while maintaining operational simplicity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables more precise defect determination by considering multiple features and reducing misjudgment, as it compares feature images rather than relying solely on global feature gray values, enhancing the accuracy of defect detection in charged particle beam imaging.
Implementation Method 1
Charged particle microscopic images are formed by detecting charged particles released from a sample being bombarded by a charged particle beam
Data Source
AI summary
A method for determining a defect during sample inspection involving charged particle beam imaging transforms a target charged particle microscopic image and its corresponding reference charged particle microscopic images each into a plurality of feature images, and then compares the feature images against each other. Each feature image captures and stresses a specific feature which is common to both the target and reference images. The feature images produced by the same operator are corresponding to each other. A distance between corresponding feature images is evaluated. Comparison between the target and reference images is made based on the evaluated distances to determine the presence of a defect within the target charged particle microscopic image.