Distortion Map Alignment for Semiconductor Inspection Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for aligning distorted images in lithographic device manufacturing are computationally expensive and inefficient, particularly in semiconductor substrate inspection, leading to inaccuracies in edge placement error and critical dimension measurements due to field of view distortions and charging artefacts.
Innovation Solution
An encoder-decoder network is used to determine optimized weightings for aligning distorted images by encoding and decoding reference and distorted images into a latent space, generating a distortion map that corrects for image distortions, thereby improving measurement accuracy and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional alignment methods are used to correct image distortions, then measurement accuracy can be improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The system pre-calculates and stores distortion maps for various known distortion conditions in a lookup table. During actual measurement, the system simply queries this pre-computed table rather than performing complex real-time calculations, thus achieving high measurement accuracy without computational complexity during operation
Solution Approach 2:
The patent creates simplified representations of distortion patterns through distortion maps that capture the essential transformation information. These maps serve as compact copies of complex distortion models, enabling fast application of corrections without processing the full complexity of original distortion physics
2Measurement precision
If complex alignment algorithms are applied to correct distortions, then image alignment accuracy improves, but processing speed decreases
Solution Approach 1:
Distortion correction parameters and transformation maps are computed in advance during system calibration or offline processing. The pre-computed maps are stored and directly applied during production measurements, eliminating the need for complex real-time calculations and achieving both high accuracy and fast processing
Solution Approach 2:
The system dynamically selects from multiple pre-computed distortion maps based on the specific imaging conditions and distortion characteristics observed. This allows the system to adapt to different scenarios without performing complex calculations, maintaining both accuracy and speed by choosing the most appropriate pre-computed correction
3Reliability
If traditional distortion correction methods are used, then measurement reliability can be improved, but the system becomes less adaptable to different distortion types
Solution Approach 1:
The distortion map framework provides a universal representation that can model various types of distortions (optical distortions, charging artifacts, field of view variations) through a common mathematical structure. The same correction mechanism adapts to different distortion types by using appropriately computed distortion maps, ensuring both reliability and versatility
Solution Approach 2:
The system handles different distortion types by changing the parameters within the distortion map representation rather than using fundamentally different correction methods. By adjusting the transformation parameters and map characteristics, the system reliably corrects various distortion types through a unified approach
Data Source
AI summary
A method for determining an optimized weighting of an encoder and decoder network; the method comprising:for each of a plurality of test weightings, performing the following steps with the encoder and decoder operating using the test weighting:(a) encoding, using the encoder, a reference image and a distorted image into a latent space to form an encoding;(b) decoding the encoding, using the decoder, to form a distortion map indicative of a difference between the reference image and a distorted image;(c) spatially transforming the distorted image by the distortion map to obtain an aligned image;(d) comparing the aligned image to the reference image to obtain a similarity metric; and(e) determining a loss function which is at least partially defined by the similarity metric;wherein the optimized weighting is determined to be the test weighting which has an optimized loss function.