Tomographic Reconstruction Segmentation Using Iterative Physical Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tomographic reconstruction methods primarily rely on image characteristics for segmentation, which can lead to inaccurate material characterization due to reliance on voxel intensity variations and local differences, rather than physical properties, resulting in inconsistent segmentation and labeling.
Innovation Solution
A tomographic reconstruction and segmentation process that updates atomic number (Z) and density (rho) for volume elements at multiple iterations, using a materials library for voxel matching and probabilistic classification to determine the most likely material, allowing for iterative correction of labeling errors and improved material characterization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If segmentation is based on image characteristics (voxel intensity variations and local differences), then the segmentation process is simple and fast, but the material characterization accuracy deteriorates
Solution Approach 1:
The patent changes the basis of segmentation from image characteristics (intensity variations) to physical parameters (atomic number Z and density rho). By updating these physical parameters iteratively during reconstruction and using them as the basis for segmentation, the method achieves both accurate material characterization and consistent segmentation results.
Solution Approach 2:
The patent implements a feedback mechanism where segmentation results based on physical parameters are used to update the reconstruction in subsequent iterations. The forward model verifies the integrity of segmented data by checking consistency with acquisition data, and incorrect segmentations are corrected in following iterations, improving both accuracy and reliability.
2Productivity
If segmentation is performed after one iteration of reconstruction based on voxel characteristics, then the process is efficient, but segmentation consistency deteriorates
Solution Approach 1:
The patent performs segmentation at multiple iterations rather than just once, maintaining continuous refinement of segmentation based on updated physical parameters. This continuous action ensures that segmentation results converge to consistent and accurate material characterization throughout the reconstruction process.
Solution Approach 2:
The forward model provides feedback by verifying segmentation integrity against acquisition data at each iteration. This feedback mechanism detects and corrects inconsistent segmentations, ensuring stability and consistency of the final segmentation results.
3Measurement precision
If dual-energy iterative reconstruction/segmentation is used to obtain multiple sub-sinograms and tomograms, then quantitative analysis capability is improved, but device complexity increases
Solution Approach 1:
The patent uses a unified forward model that handles both reconstruction and segmentation verification functions. This multi-functional approach allows the same model to perform quantitative analysis, verify segmentation integrity, and correct errors, reducing overall system complexity despite the advanced capabilities provided.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides accurate material characterization by segmenting based on physical properties, improving the consistency and accuracy of material labeling, and correcting for errors in subsequent iterations, leading to enhanced 3D rendering and quantitative analysis.
Implementation Method 1
If the source of radiation is outside the subject, such as in x-ray computed tomography (CT) the technique is referred to as transmission tomography
Implementation Method 2
CT is used in medical imaging, as well as geological sample, such as fossils, reservoir rocks, and soil
Data Source
AI summary
Some embodiments of the invention provide a method of determining a material characteristic of material in a sample by iterative tomographic reconstruction. The method conducts one or more X-ray tomography scans of a sample, and then determines one or more estimated material characteristics, such as atomic number and density, for multiple volume elements in the sample using a tomographic reconstruction algorithm. These estimated material characteristics are then modified by reference to stored known material characteristic data. Preferably, determining the composition of the sample volume during reconstruction includes segmenting the sample into regions of common composition, the segmenting being performed during iterative reconstruction instead of being based on the voxel characteristics determined upon the completion of iterative reconstruction. Preferred versions will perform one or more additional iterations of the tomographic reconstruction algorithm, where each iteration updates the one or more estimated material characteristics for the volume elements.