Lamella Structure Selection for Automated Cryo Milling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for preparing cryogenic lamella samples require manual user interaction, leading to inefficiencies such as increased storage time, higher costs, and sample degradation due to limited workflow capacity and susceptibility to errors.
Innovation Solution
An automated system using machine learning models for image segmentation and structure-of-interest selection, which applies automated image processing to identify and select suitable structures for milling in cryogenic lamella samples, thereby setting appropriate milling parameters for scientific instruments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual user interaction is used for structure selection, then ease of operation is maintained, but productivity is reduced and loss of time increases
Solution Approach 1:
The system performs automated image processing and structure selection without requiring manual user interaction. The machine learning model independently segments images, identifies candidate structures, and selects optimal structures for milling, allowing the system to serve itself rather than relying on human operators for each step.
Solution Approach 2:
The patent replaces the mechanical/manual system of human operators visually selecting structures with an automated computational system using machine learning models. The image processing and structure selection functions are transferred from human operators to automated algorithms, significantly increasing throughput and eliminating manual bottlenecks.
2Adaptability or versatility
If manual user interaction is used for structure selection, then adaptability to complex samples is maintained, but loss of time increases and reliability decreases
Solution Approach 1:
The system incorporates feedback mechanisms where the machine learning model continuously processes images and refines structure selections based on segmented classes and candidate evaluations. The automated workflow provides consistent feedback loops for image processing, candidate identification, and final selection, reducing human error while maintaining adaptability through algorithmic adjustments.
3Productivity
If automated image processing is implemented, then productivity increases and loss of time decreases, but device complexity increases
Solution Approach 1:
The automated system segments the image processing task into distinct computational stages: initial image segmentation into classes, identification of candidate structures from segmented classes, and final selection of optimal structures. This segmentation of the automated workflow manages complexity by breaking down the overall process into manageable, modular steps that can be executed sequentially by the machine learning model.
Data Source
Figure 1~2
Figure 3
Figure 4A~4B
AI summary
Disclosed herein are scientific instrument support systems, as well as related methods, computing devices, and computer-readable media. For example, in some embodiments, a support apparatus is provided for a charged particle microscope. The support apparatus is configured to apply automated image processing to an image representing a lamella sample to segment the image into a plurality of segmented classes. The support apparatus is also configured to identify, based on the plurality of segmented classes, a subset of candidate structures-of-interest in the lamella sample and to select, from the subset of candidate structures-of-interest in the lamella sample, a selected structure-of-interest for milling. The support apparatus is also configured to set, based on the selected structure-of-interest for milling, at least one milling parameter for the scientific instrument. An automated method performed via a computing device for providing such scientific instrument support is also provided.