Scanning Microscope Partial-Region Imaging With ML Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing microscope systems require manual determination of scan patterns and reconstruction of images, which is time-consuming and lacks automation, and does not optimally incorporate context information for image reconstruction.
Innovation Solution
A method using a machine learning system trained through supervised, unsupervised, and reinforcement learning to automatically determine scan patterns and reconstruct overall images from scanned partial regions of a sample in a scanning microscope.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If manual determination of scan patterns and reconstruction of images is used, then expertise and context understanding are applied, but the process is time-consuming and lacks automation
Solution Approach 1:
The machine learning system enables the microscope to automatically determine optimal scan patterns and reconstruct images without human intervention. The system processes raw scan data, identifies relevant features, and generates complete images autonomously, making the system self-sufficient in tasks that previously required expert operators.
Solution Approach 2:
The patent replaces manual expert analysis with an automated machine learning-based image reconstruction system. The ML model processes scan data and reconstructs images algorithmically, substituting the mechanical process of human expert work with an automated computational system that operates without human intervention.
2Productivity
If sparse scanning of partial regions is used, then recording speed increases and radiation exposure decreases, but image reconstruction quality may be compromised
Solution Approach 1:
The machine learning system incorporates feedback mechanisms where the reconstruction quality is continuously evaluated and used to adjust scanning parameters. The system learns from reconstruction outcomes and refines its scan pattern selection to optimize both speed and quality, using feedback loops to balance sparse scanning with image fidelity.
Solution Approach 2:
The patent dynamically adjusts scanning parameters such as scan density, region selection, and sampling rates based on image characteristics and reconstruction requirements. The machine learning system modifies these parameters in real-time to achieve optimal balance between recording speed and image quality for different sample types and imaging conditions.
3Device complexity
If traditional optimization methods are used for image reconstruction, then computational simplicity is maintained, but context information is not incorporated and reconstruction is not optimal
Solution Approach 1:
The patent introduces a machine learning model as an intermediary between raw scan data and final image reconstruction. This ML intermediary processes the sparse scan data, incorporates contextual information about sample characteristics and imaging conditions, and generates high-quality reconstructions that bridge the gap between simple algorithms and optimal results.
Data Source
AI summary
A method is useful for scanning partial regions of a sample by a scanning microscope, such as a laser scanning microscope or a scanning electron microscope, and for reconstructing an overall image of the sample from data of the scanned partial regions of the sample. The method includes: 1) determining partial regions of the sample, which are scanned by the scanning microscope, by a machine learning system which is trained by supervised learning, unsupervised learning, and/or reinforcement learning for improved determination of the partial regions of the sample which are scanned by the scanning microscope; 2) scanning the determined partial regions of the sample by the scanning microscope; and 3) reconstructing the overall image of the sample from the data of the scanned partial regions of the sample, wherein non-scanned partial regions of the sample are estimated by the data of the scanned partial regions of the sample.