Wafer Defect Classification From Preclustered Normal Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor processing technologies face challenges in accurately detecting and classifying defects on wafers, particularly in ensuring high throughput and quality while minimizing costs, with existing methods like visual and electrical inspections being inadequate.
Innovation Solution
A method involving image processing using a trained machine learning model, where images are divided into patches, feature groups are clustered, and a target feature group is selected based on distance and threshold values to determine defect presence and type, utilizing augmented images for training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional visual and electrical inspection methods are used, then the inspection process is simple, but the defect detection accuracy is insufficient
Solution Approach 1:
The wafer image is divided into multiple patches, and each patch is processed independently through feature extraction and clustering. This segmentation allows the complex inspection task to be broken down into manageable parts, improving detection accuracy while maintaining systematic control over the overall process
Solution Approach 2:
Feature groups are pre-clustered from multiple second images before the actual defect detection on the target wafer image. This preliminary clustering creates a reference database of normal and defective patterns, enabling more accurate and efficient defect identification when the target image is analyzed
2Measurement precision
If machine learning models are used for defect detection, then the defect classification accuracy is improved, but the processing time increases
Solution Approach 1:
By dividing the wafer image into patches and processing them in parallel through the machine learning model, the system achieves accurate defect classification while reducing overall processing time compared to analyzing the entire image sequentially
Solution Approach 2:
The system extracts only the most relevant features from each patch using clustering algorithms, rather than processing all possible image data. This selective feature extraction maintains high classification accuracy while significantly reducing the computational burden and processing time
3Reliability
If multiple feature groups are clustered from multiple images, then the defect detection reliability is improved, but the computational complexity increases
Solution Approach 1:
Multiple second images are processed in advance to pre-cluster feature groups that represent normal and defective patterns. This preliminary action creates a reusable reference database that improves detection reliability across multiple wafers without requiring the same computationally intensive clustering process to be repeated for each new wafer
Solution Approach 2:
The system transforms raw image data into clustered feature groups with optimized parameters. By changing the representation from raw pixels to clustered feature vectors, the system maintains high detection reliability while reducing the dimensionality and complexity of the data that needs to be processed for each wafer
Data Source
AI summary
An electronic device for detecting and classifying a defect of a wafer and a method of detecting and classifying a defect of a wafer are provided. The method includes obtaining an image of a wafer, dividing the image into a plurality of patches and extracting first features from the plurality of patches, selecting a target feature group from a plurality of feature groups, based on the first features, determining whether a defect exists in the wafer, based on the target feature group and the first features, and when the defect exists in the wafer, determining a defect type of the defect, wherein the plurality of feature groups is obtained through clustering based on second features extracted from a plurality of normal images.