SEM Image Denoising and Alignment for Clearer Microscopy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scanning electron microscope (SEM) images suffer from noise that degrades image quality and resolution, particularly in applications like semiconductor process evaluation.
Innovation Solution
A method involving obtaining multiple SEM images, removing noise from each using machine learning (e.g., auto-encoders/decoders), aligning these images to minimize mean squared error, and combining them to generate a high-quality SEM image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple SEM images are captured and combined, then image quality and resolution are improved, but processing time and computational complexity increase
Solution Approach 1:
The patent applies preliminary action by performing noise removal using machine learning on individual SEM images before alignment and combination. This preprocessing step ensures that noise is eliminated early in the workflow, preventing it from propagating through subsequent processing stages and reducing the need for repeated processing iterations, thereby optimizing the overall processing time while maintaining image quality.
Solution Approach 2:
The patent segments the image processing workflow into distinct stages: noise removal, alignment, and combination. By dividing the processing of multiple SEM images into these separate sequential steps, the system can optimize each stage independently and process images more efficiently, reducing total processing time while achieving superior image quality through systematic handling of each processing aspect.
2Measurement precision
If noise removal is applied to each SEM image, then image clarity is improved, but computational complexity and processing resources increase
Solution Approach 1:
The patent replaces traditional mechanical or manual noise filtering methods with machine learning-based noise removal. This substitution enables automated, adaptive noise elimination that achieves superior image clarity while optimizing computational resource usage through learned patterns rather than exhaustive processing, thereby reducing the perceived computational complexity despite the advanced algorithms employed.
3Speed
If alignment processing is performed on noisy images, then processing speed is maintained, but image blur and misalignment occur
Solution Approach 1:
The patent applies preliminary noise removal before the alignment process, ensuring that images are cleaned of noise prior to alignment operations. This sequencing prevents noise from causing misalignment or blur during the alignment stage, achieving high alignment accuracy while maintaining efficient processing speed through the optimized workflow arrangement.
Data Source
AI summary
Provided is a scanning electron microscope (SEM) image improving method including obtaining a plurality of first SEM images by using the SEM, generating a plurality of second SEM images by removing noise from each of the first SEM images, aligning the plurality of second SEM images, and generating a third SEM image by combining the aligned plurality of second SEM images.