System and method using multi-modality imaging to estimate effective medium parameters
AssigneesUniversity of Central Florida Research Foundation, Inc., iTomography Corporation
InventorsAlexander Katsevich, Michael Frenkel, Igor Frenkel, Airidas Korolkovas, Seongjin Yoon
Life of this patent
Estimating the term…
Abstract
In various embodiments, the present invention provides a multi-modality imaging system in combination with a novel imaging method comprising complex modeling of the modalities and machine learning. In a particular embodiment, multi-modality imaging of a large rock core is described.
Claims29
A computer implemented method for training a machine learning prediction model for performing multi-modality non-destructive tomographic imaging of an object of interest, the method comprising:
- providing a plurality of models of an object of interest, each of the plurality of models comprising one or more predetermined material properties;
- computing an X-ray Computed Tomography (CT) reconstruction for each of the plurality of models on a 3D grid by computing one or more effective material properties of the models that a selected X-ray CT scanning instrument is sensitive to for each of the plurality of models with a 3D grid step size corresponding to the selected X-ray CT scanning instrument resolution;
- computing an Electromagnetic (EM) reconstruction for each of the plurality of models on a 3D grid by computing one or more effective material properties of the models that a selected EM Tomography scanning instrument is sensitive to for each of the plurality of models with a 3D grid step size corresponding to the selected EM Tomography scanning instrument resolution;
- and
- training a machine learning prediction model to estimate a value of one or more of the material properties for each of the plurality of models on a 3D grid having a desired 3D grid step size, wherein the estimate of the value of the one or more material properties is determined based upon the computed X-ray CT reconstruction and computed EM reconstruction of each of the plurality of models.
A computer implemented method for performing multi-modality non-destructive tomographic imaging of one or more objects of interest, the method comprising:
- performing an X-ray Computer Tomography (CT) scan of one or more objects of interest using a selected X-ray CT instrument to generate X-ray CT scan data of the one or more objects of interest;
- performing an EM Tomography scan of the one or more objects of interest using a selected EM Tomography scanning instrument to generate EM Tomography scan data of the one or more objects of interest;
- reconstructing one or more effective material properties of the one or more objects of interest that the selected X-ray CT scanning instrument is sensitive to by inverting the X-ray CT scan data on a 3D grid having a 3D grid step size corresponding to a resolution of the selected X-ray CT instrument;
- reconstructing one or more effective material properties of the one or more objects of interest that the selected EM Tomography scanning instrument is sensitive to by inverting the EM Tomography scan data on a 3D grid having a 3D grid step size corresponding to a resolution of the selected EM Tomography scanning instrument resolution;
- and
- using a machine learning prediction model to estimate a value of one or more material properties for the one or more objects of interest on a 3D grid having a desired 3D grid step size, wherein the machine learning prediction model comprises an empirical function to convert the one or more effective material properties of the one or more objects of interest that the selected X-ray CT scanning instrument is sensitive to and the one or more effective material properties of the one or more objects of interest that the selected EM Tomography scanning instrument is sensitive to into the estimate of the value of the one or more material properties that both the selected X-ray CT scanning instrument and the selected EM Tomography scanning instrument are not sensitive to.
One or more non-transitory computer-readable media having computer-executable instructions for performing computer-executable instructions for performing a method of running a software program on a computing device for training a machine learning prediction model for providing a multi-modality non-destructive tomographic imaging method for one or more objects of interest, the method including issuing instructions from the software program comprising:
- providing a plurality of models of an object of interest, each of the plurality of models comprising one or more predetermined material properties;
- computing an X-ray Computed Tomography (CT) reconstruction for each of the plurality of models on a 3D grid by computing one or more effective material properties of the models that a selected X-ray CT scanning instrument is sensitive to for each of the plurality of models with a 3D grid step size corresponding to the selected X-ray instrument resolution;
- computing an Electromagnetic (EM) reconstruction for each of the plurality of models on a 3D grid by computing one or more effective material properties of the models that a selected EM Tomography scanning instrument is sensitive to for each of the plurality of models with a 3D grid step size corresponding to the selected EM Tomography scanning instrument resolution;
- and
- training a machine learning prediction model to estimate a value of one or more of the material properties for each of the plurality of models on a 3D grid having a desired 3D grid step size, wherein the estimate of the value of the one or more material properties is determined based upon the computed X-ray CT reconstruction and computed EM reconstruction of each of the plurality of models.
Dependencies are inferred from the claim text.