Defect Management Apparatus Label Diversion for Semiconductor Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, defect management is hindered by the labor-intensive process of preparing large numbers of labels for machine learning classification models, especially in high-volume production environments where time constraints limit the resolution of inspection images and the accuracy of defect classification.
Innovation Solution
A defect management apparatus that acquires and maps inspection data from multiple devices, automatically diverting classification labels from one inspection device to another, enabling efficient generation of proper labels and reducing manual labor through a system comprising an acquisition unit, mapping unit, and label diversion unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual classification labeling is performed to train machine learning models, then classification accuracy can be improved, but the time and labor required increases significantly
Solution Approach 1:
The system copies classification labels from inspection devices with high resolution to inspection devices with low resolution. The label diversion unit transfers labels generated by the first inspection device to the second inspection device, eliminating the need for manual labeling at the second device while maintaining classification accuracy.
Solution Approach 2:
The system enables self-service by allowing inspection devices to automatically generate and transfer their own classification labels. The label diversion unit automatically diverts labels from the first inspection device to the second inspection device without requiring manual intervention, reducing both time and labor costs.
2Measurement precision
If high-resolution inspection images are used to improve defect detection accuracy, then measurement precision improves, but processing time and resource requirements increase
Solution Approach 1:
The system merges the strengths of multiple inspection devices by combining high-resolution defect detection capabilities with low-resolution rapid inspection capabilities. The label diversion unit integrates labels from high-resolution inspection to enhance the performance of low-resolution inspection, achieving both accuracy and efficiency.
Solution Approach 2:
The system applies local quality by using high-resolution inspection and labeling only for specific critical defect types, while using low-resolution inspection for general screening. The label diversion unit selectively transfers labels based on defect characteristics, optimizing resource allocation and processing time.
3Productivity
If multiple inspection devices with different resolutions are used to handle high-volume production, then productivity increases, but consistency in defect classification decreases
Solution Approach 1:
The system achieves universality by creating a unified classification framework that works across multiple inspection devices with different resolutions. The label diversion unit establishes a common labeling standard that can be applied universally to all devices, ensuring consistent defect classification regardless of the specific inspection device used.
Solution Approach 2:
The label diversion unit acts as an intermediary that mediates between inspection devices with different resolutions. It transfers and standardizes labels from high-resolution devices to low-resolution devices, ensuring classification consistency across the entire inspection system while maintaining high productivity.
Data Source
AI summary
According to one embodiment, a defect management apparatus includes a processor. The processor acquires first information and second information, the first information including first defect positions relating to defects detected with a first device for an inspection target and corresponding first labels indicating classifications of the defects, the second information including second defect positions relating to defects detected with a second device for the inspection target. The processor determines a first defect position corresponding to a second defect position as a corresponding defect position. The processor diverts the first label corresponding to the corresponding defect position as a second label of the second defect position.