Automated Dimension Measurement for Semiconductor Cross-Sectional Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current dimension measurement methods for semiconductor devices, particularly in cross-sectional SEM and TEM images, face challenges such as operator dependence, difficulty in extracting outlines due to white shadows, and inconsistency in measurement results due to varying training data sets and models, leading to inaccurate and time-consuming manual measurements.
Innovation Solution
A dimension measuring device that employs machine learning and image processing techniques to automatically extract region boundaries and feature points from cross-sectional SEM images, reducing operator error and measurement time by using a combination of region division units and a dimension measuring unit that calculates coordinates based on these features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual measurement is applied to advanced semiconductor processes, then measurement can be performed on complex structures, but operator dependence and human error increase measurement variability
Solution Approach 1:
The measurement system performs automatic dimension extraction and measurement without human intervention. The computer automatically processes images, extracts dimensions, and generates measurement results, eliminating operator dependence and human error while maintaining measurement accuracy on complex semiconductor structures
Solution Approach 2:
The patent replaces manual mechanical measurement operations with automated image processing and computer-based dimension extraction algorithms. This substitution eliminates the mechanical aspects of manual measurement while preserving and enhancing measurement precision through consistent automated processing
2Measurement precision
If each individual pattern is measured one by one in manual measurement, then detailed measurement can be achieved, but measurement time increases significantly
Solution Approach 1:
The system merges multiple individual pattern measurements into a single automated processing operation. By processing all patterns in an image simultaneously through automated dimension extraction, the system maintains detailed measurement capability while dramatically reducing total measurement time
Solution Approach 2:
The automated measurement system performs continuous processing of multiple patterns without interruption or manual repositioning. The computer continuously extracts dimensions from all patterns in the image, eliminating the stop-start nature of manual measurement and maintaining high productivity throughout the measurement process
3Measurement precision
If re-measurement is performed when new measurement places are identified, then measurement accuracy can be improved, but time consumption increases
Solution Approach 1:
The automated measurement system performs preliminary extraction of dimensions from all possible measurement locations in a single processing pass. When new measurement places are identified during optimization, the data is already available in the extracted dimension information, eliminating the need for time-consuming re-measurement while maintaining measurement accuracy
4Illumination intensity
If white shadows occur at interfaces in cross-sectional SEM images, then image quality is affected, but outline extraction becomes difficult
Solution Approach 1:
The patent introduces image processing algorithms as an intermediary between the raw SEM image and the final dimension extraction. These algorithms process the image data to compensate for white shadow effects, enhancing outline detectability without requiring changes to the illumination conditions that cause the shadows
Solution Approach 2:
The system changes the processing parameters of the image data through automated image processing and enhancement techniques. By adjusting contrast, applying edge detection algorithms, and modifying how the image data is interpreted, the system overcomes the measurement difficulties caused by white shadows while preserving the original image quality
Data Source
AI summary
The present disclosure relates to a dimension measuring device that shortens a time required for dimension measurement and eliminates errors caused by an operator. A dimension measuring device that measures a dimension of a measurement target using an input image is provided, in which a first image in which each region of the input image is labeled by region is generated by machine learning, an intermediate image including a marker indicating each region of the first image is generated based on the generated first image, a second image in which each region of the input image is labeled by region is generated based on the input image and the generated intermediate image, coordinates of a boundary line between adjacent regions are obtained by using the generated second image, coordinates of a feature point that defines a dimension condition of the measurement target are obtained by using the obtained coordinates of the boundary line, and the dimension of the measurement target is measured by using the obtained coordinates of the feature point.