Wafer Mark Identification Using AI and Transmissive Illumination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chemical residues often obscure wafer marks in the semiconductor industry, making it difficult for existing systems to accurately identify and process these marks, which is crucial for semiconductor device control and management.
Innovation Solution
An image processing system that includes a camera, a light source capable of projecting through chemical residues, and an AI statistical model for identifying wafer marks, combined with an advanced image processing engine using edge detection and image stacking techniques to enhance recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image recognition methods are used, then the system is simple and fast, but it cannot accurately identify wafer marks obscured by chemical residues
Solution Approach 1:
The patent introduces an AI statistical model as an intermediary between the image capture system and the identification output. This model acts as a mediator that processes the raw image data and probabilistic information to achieve accurate identification of obscured wafer marks, resolving the contradiction by adding intelligence without excessive complexity
Solution Approach 2:
The system changes the parameter of image processing from traditional deterministic methods to probabilistic AI-based methods. By using statistical models that output probability distributions rather than fixed thresholds, the system can handle the variability and obscuration caused by chemical residues, improving measurement precision
2Measurement precision
If multiple images are captured and processed using statistical analysis, then the identification accuracy of obscured characters improves, but the processing time increases
Solution Approach 1:
The system performs preliminary actions by capturing multiple images and pre-processing them through the AI statistical model before final identification. By preparing probabilistic data in advance and using edge detection to pre-identify potential character locations, the system reduces the computational burden during final decision-making, balancing accuracy with processing time
Solution Approach 2:
The patent segments the image processing task into distinct stages: edge detection to identify potential character regions, AI statistical model processing to generate probability distributions, and threshold-based final identification. This segmentation allows each stage to be optimized independently, improving overall efficiency while maintaining high accuracy
3Illumination intensity
If a light source is used to project through chemical residues, then the visibility of wafer marks improves, but the system complexity increases
Solution Approach 1:
The light source system is designed to serve multiple functions: it provides illumination for image capture, enables projection through chemical residues, and works with the AI model to enhance visibility. By making the optical system multi-functional rather than adding separate dedicated systems, the patent improves wafer mark visibility while minimizing the increase in overall system complexity
Data Source
AI summary
The present disclosure provides an image processing system and method. The system includes an image capturing unit and an image processing unit. The image capturing unit includes a camera, a light source, and a wafer transfer stage. The camera captures an image of a wafer mark. The light source is capable of projecting through a plurality of chemical residues on the wafer mark. The image processing unit includes an image processor and an object matching subsystem. The image processor is configured to perform calculations of the image processing unit to generate an output result. The object matching subsystem includes an object matching unit and a memory unit. The object matching unit includes an artificial intelligence (AI) statistical model for identifying each character of the wafer mark. The memory unit stores a set of probabilities of the object matching unit detecting each character of the wafer mark in each image.