Integrated Circuit Pattern Defect Detection With Transformation Fields
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect detection methods for integrated circuit patterns, such as photolithography masks and wafers, are computationally expensive, time-consuming, and require extensive user effort, limiting their throughput and accuracy in quality assurance processes.
Innovation Solution
A computer-implemented method that utilizes a transformation field pair to register imaging and reference datasets directly, without warping, and employs machine learning models to detect defects, reducing computation time and improving accuracy by using pre-registered datasets and probabilistic generative models to measure registration uncertainty.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional defect detection methods are used, then defect detection capability is provided, but computation time and user effort are excessive
Solution Approach 1:
The patent applies preliminary action by pre-registering the imaging dataset to the reference dataset before defect detection. This pre-registration step establishes a transformation field pair that maps the imaging dataset to the reference dataset's coordinate system, eliminating the need for time-consuming registration during the actual defect detection process. The pre-computed transformation fields enable direct comparison of pixel values without real-time registration computations.
2Measurement precision
If traditional defect detection methods are used, then defect detection capability is provided, but throughput is limited
Solution Approach 1:
The patent performs pre-registration of the imaging dataset to the reference dataset before defect detection, computing transformation field pairs in advance. This preliminary computation eliminates the need for time-consuming registration during defect detection, enabling direct pixel-value comparison and significantly increasing throughput while maintaining detection accuracy.
Solution Approach 2:
The patent replaces the mechanical registration process (which involves iterative alignment and warping computations) with a substitution approach using pre-computed transformation field pairs. These transformation fields directly map coordinate systems without requiring iterative geometric transformations, substituting complex computational mechanics with more efficient linear or affine transformations that enable high-throughput processing.
3Measurement precision
If extensive user effort is required, then defect detection can be performed, but efficiency is reduced
Solution Approach 1:
The patent implements self-service by automatically performing the registration process through pre-computed transformation field pairs. The system autonomously aligns the imaging dataset with the reference dataset using the pre-established transformation fields, eliminating the need for manual intervention in the registration step. Users only need to provide the imaging dataset and reference dataset, and the system automatically handles the complex alignment and defect detection processes.
Data Source
AI summary
The invention relates to a computer implemented method for defect detection comprising: obtaining an imaging dataset of an object comprising integrated circuit patterns; obtaining a reference dataset of the object; registering the imaging dataset and the reference dataset by obtaining at least one transformation field pair comprising an input transformation field and a corresponding reference transformation field, wherein the input transformation field or the reference transformation field can be zero; and detecting defects in the imaging dataset using the at least one obtained transformation field pair. The invention also relates to a computer-readable medium, a computer program product and a system for detecting defects.