Zernike Polynomial Substrate Reference Image Creation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional defect inspection methods in semiconductor manufacturing are prone to false defects due to color unevenness, and the creation of a substrate reference image can vary based on operator skill, leading to inconsistent inspection results.
Innovation Solution
The method involves decomposing substrate images using Zernike polynomials to calculate coefficients, extracting median and deviated values, and combining specified images to create a reference image with filtering capabilities against color unevenness, ensuring consistent defect inspection regardless of operator skill.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If substrate images with color unevenness are combined to create a reference image, then false defects are filtered out, but the reference image creation becomes subjective and inconsistent
Solution Approach 1:
The patent transforms the subjective image selection process into an objective mathematical process by changing parameters from operator judgment to Zernike polynomial coefficients. By decomposing images into standardized mathematical components and using automated coefficient comparison, the system eliminates operator skill variability while maintaining the ability to filter false defects through systematic selection of images with representative color unevenness patterns.
Solution Approach 2:
The patent replaces the mechanical/manual process of operator-based image selection with an automated computational system. Instead of operators manually selecting and combining images based on empirical rules, the system uses Zernike polynomial decomposition and automated coefficient analysis to objectively select and combine images, substituting human judgment with mathematical algorithms.
2Stability of the object's composition
If operator skill level varies, then reference image creation quality varies, but automation increases device complexity
Solution Approach 1:
The patent introduces Zernike polynomials as an intermediary mathematical framework that bridges the gap between simple image combination and complex automated processing. The polynomials serve as a standardized intermediate representation that enables objective comparison and selection of substrate images through coefficient analysis, providing a systematic mediator between the input images and the final reference image.
Solution Approach 2:
The patent segments the image processing task into distinct mathematical components using Zernike polynomial decomposition. By breaking down the complex image data into standardized polynomial coefficients representing different spatial frequencies and patterns, the system enables systematic and automated selection criteria, transforming a holistic subjective judgment into separable objective measurements.
Data Source
AI summary
In the present invention, a planar distribution of pixel values in a picked-up substrate image is decomposed into a plurality of pixel value distribution components through use of a Zernike polynomial for each of substrate images; Zernike coefficients of the pixel value distribution components decomposed through use of the Zernike polynomial are calculated; a median value and values deviated from the median value by a predetermined value or more are extracted for every Zernike coefficients having a same couple of degrees from the calculated Zernike coefficients; substrate images having the extracted values are specified; and a substrate image being a defect inspection reference is created by combining the specified substrate images.