Image Alignment for Noisy Semiconductor Specimens
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image alignment methods in semiconductor quality control, such as normalized cross-correlation, fail to accurately align noisy semiconductor images with repetitive patterned features, leading to potential alignment errors due to noise influence and lack of distinguishable horizontal or vertical features.
Innovation Solution
The system employs denoising techniques like structure-based denoising and singular value decomposition (SVD) to reduce noise and detect horizontal or vertical features, designating them as a region of interest (ROI) for improved alignment using normalized cross-correlation (NCC) or other alignment methods, achieving sub-pixel accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional normalized cross-correlation alignment is used on noisy images, then the alignment process is simple, but the alignment accuracy deteriorates due to noise influence and lack of distinguishable features
Solution Approach 1:
The patent applies denoising processing to images before performing alignment operations. By pre-processing the images to remove noise and enhance features, the alignment algorithm can achieve higher accuracy without being misled by noise patterns. This preliminary action prepares the data in advance to ensure successful alignment despite initial poor image quality.
Solution Approach 2:
The patent segments the image processing into distinct stages: denoising, feature enhancement, and alignment. By dividing the complex task into separate processing steps, each stage can be optimized independently, allowing high alignment accuracy to be achieved through a systematic multi-step approach rather than a single complex operation.
2Measurement precision
If alignment is performed on images with repetitive patterned features, then coverage is complete, but alignment accuracy deteriorates due to inability to distinguish unique features
Solution Approach 1:
The patent enhances local features within the image to make repetitive patterns more distinguishable. By applying denoising and feature enhancement techniques locally across the image, unique characteristics are preserved or amplified in each region, enabling the alignment algorithm to identify distinguishable features even within repetitive patterns.
3Measurement precision
If noise reduction is applied to images, then alignment accuracy improves, but processing time increases
Solution Approach 1:
The patent applies denoising processing selectively to regions of the image that contain alignment-critical features, rather than uniformly processing the entire image. This partial action approach reduces the computational burden and processing time while still achieving sufficient noise reduction in the areas that matter most for alignment accuracy.
Data Source
AI summary
Methods and systems for aligning images of a specimen are provided. One method includes reducing noise in a test image generated for a specimen by an imaging subsystem thereby generating a denoised test image. The method also includes detecting one or more patterned features in the denoised test image extending in at least a horizontal or vertical direction. In addition, the method includes designating an area of the denoised test image in which the detected one or more patterned features are located as a region of interest in the denoised test image. The method further includes aligning the denoised test image to a reference image for the specimen using only the region of interest in the denoised test image and a corresponding area in the reference image.