Wafer Spatial Pattern Classification via Neural Network Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying and classifying wafer spatial patterns in semiconductor manufacturing are inefficient and prone to human error, as they rely on manual analysis, leading to increased workload and potential neglect of defects.
Innovation Solution
A classification method that performs statistical analysis using the odds ratio hypothesis-testing method, followed by denoising and signal enhancement via cellular automata, and feature extraction using a neural network, culminating in automated wafer spatial pattern classification through clustering algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual analysis is used to identify wafer spatial patterns, then engineers can directly judge defect distributions, but the workload increases and efficiency decreases
Solution Approach 1:
The patent replaces the manual mechanical analysis process with an automated image processing system. The system uses computer algorithms to perform statistical analysis, denoising, signal enhancement, feature extraction, and clustering classification on wafer defect images, substituting human engineers' manual inspection with automated computational methods that maintain accuracy while dramatically improving throughput
Solution Approach 2:
The patent implements a self-service automated classification system where the wafer defect analysis is performed autonomously without human intervention. The system automatically processes defect images through multiple computational stages including statistical analysis, cellular automata-based denoising, neural network feature extraction, and clustering classification, enabling the system to serve itself in identifying and categorizing spatial patterns
2Reliability
If manual screening of wafer patterns is performed to ensure no cases are neglected, then comprehensive defect detection is achieved, but the time consumption increases significantly
Solution Approach 1:
The patent applies preliminary action through pre-trained neural network models and pre-established clustering algorithms. The system has been预先 trained on defect data to recognize patterns, enabling it to quickly classify new wafer defects without requiring time-consuming manual inspection. The automated pipeline performs statistical analysis, denoising, and feature extraction in a pre-configured sequence that ensures comprehensive detection while minimizing time loss
Solution Approach 2:
The patent replaces the time-consuming manual screening process with automated image processing and machine learning algorithms. The system uses computer vision techniques to detect and classify spatial patterns that would require extensive manual review, achieving both comprehensive defect detection and rapid processing speeds
3Productivity
If automated classification methods are implemented, then analysis efficiency improves, but the system complexity increases
Solution Approach 1:
The patent segments the complex automated classification task into distinct modular stages: statistical analysis of defect distribution, denoising and signal enhancement using cellular automata, feature extraction through neural networks, and final clustering classification. Each module handles a specific aspect of the analysis, making the overall complex system manageable through functional decomposition and enabling high throughput processing
Data Source
AI summary
The present invention provides a classification method for automatically identifying wafer spatial pattern distribution, comprising the following steps: performing statistical analysis to distribution of defects on a wafer, the defects being divided into random defects, repeated defects and cluster defects; performing denoising and signal enhancement to the cluster defects; performing feature extraction to the cluster defects after denoising and signal enhancement; and performing wafer spatial pattern distribution classification to the cluster defects after feature extraction. By performing statistical analysis and neural network training to a great amount of wafer defect distribution, the spatial patterns in defect distribution can be automatically identified, the automatic classification of wafer spatial patterns can be realized, the workload of engineers is effectively reduced and the tracing of the root cause of such spatial pattern is facilitated.