Cascade Defect Detection for IC Imaging Datasets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect detection methods for objects with integrated circuit patterns are inefficient, requiring high computation time and struggling to achieve high recall and precision, especially when dealing with large imaging datasets containing few defects.
Innovation Solution
A computer-implemented method involving a cascade of stage-specific defect detection methods, each applying a different approach to progressively filter out defect-free candidates, thereby reducing the dataset size efficiently and improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single high-precision defect detection method is applied to the entire imaging dataset, then detection precision is improved, but computation time increases significantly
Solution Approach 1:
The defect detection process is segmented into multiple stages (first stage, second stage, third stage), where each stage applies a different detection method with increasing precision. The first stage uses a fast method to filter obvious defects, the second stage applies a medium-precision method, and the third stage uses high-precision methods only on remaining candidates. This segmentation allows the system to achieve high overall precision while minimizing computation time by avoiding application of high-precision methods to the entire dataset.
2Reliability
If multiple defect detection methods are applied to all defect candidates, then recall is improved, but computation time and processing overhead increase
Solution Approach 1:
The first stage defect detection method is applied preliminarily to all defect candidates to identify and filter out obvious defects. This preliminary action reduces the number of candidates that need to be processed by subsequent detection methods, thereby improving recall without proportionally increasing computation time. The cascaded structure ensures that multiple methods are applied selectively rather than to all candidates.
3Productivity
If a simple defect detection method is used, then computation time is reduced, but detection precision and recall deteriorate
Solution Approach 1:
The defect detection system dynamically adapts its precision level based on the characteristics of each defect candidate. The cascaded structure allows the system to start with fast, simple detection and progressively apply more complex, precise methods only when necessary. This dynamic approach ensures that computation time is minimized for obvious cases while maintaining high precision for difficult-to-detect defects.
Data Source
AI summary
The invention relates to a computer implemented method for defect detection in an imaging dataset of an object comprising integrated circuit patterns, the method comprising: obtaining defect candidates in the imaging dataset; subsequently carrying out at least two stages, each stage comprising the following steps: applying a stage specific defect detection method to the defect candidates; discarding defect-free defect candidates; obtaining detected defects in the imaging dataset from the remaining defect candidates. The invention also relates to a corresponding computer-readable medium, computer program and system.