Material identification using multi-energy ct image data

Abstract

Disclosed are methods for identification and quantification of a number of different materials within an object using one or more multi-energy CT imaging devices and the image data sets produced therefrom. Identification and quantification of different materials is achieved by using the following three properties: solve only for sparse solutions; separate the soft tissue problem from the dense material problem; and use a combinatorial approach to allow for simple application of different constraints to different combinations of materials. Also disclosed are one or more computer program products, computer systems or computer implemented methods for the identification of multiple materials within an object.

1 . A computer-implemented method for selecting a material decomposition algorithm using image data generated using a multi-energy computer tomography (CT) imaging system using three or more energy bands using a computer program product, the method including;
 a) receiving a reconstructed or non-reconstructed multi-energy CT image data set produced using three or more energy bands in relation to an object;   b) storing said data set on a data storage medium;   c) if data received in a) is non-reconstructed, processing the data using one or more reconstruction techniques to produce a data set of reconstructed voxels;   d) comparing each voxel in the reconstructed data set to a reference set of material signal amplitude and noise properties stored on a data storage medium;   e) classifying the voxels into air, low-density or high-density voxels;   f) based on the determination of e) selecting a first material decomposition method to be applied to low-density voxels and/or selecting a second material decomposition method to be applied to higher density voxels; and   g) presenting the selected first and/or second decomposition method to an interface.   
     
     
         2 . The method of  claim 1 , wherein the method further includes the step of scanning an object using a multi-energy CT system using three or more energy bands to produce an image data set and the step of sending image data to be received at step a). 
     
     
         3 . The method of  claim 1 , wherein the step of classifying voxels as air, low-density or high-density voxels includes comparing individual voxels to a reference distribution of a specific reference material using the Mahalanobis distance metric and Euclidian distance metric. 
     
     
         4 . The method of  claim 1 , wherein the step of classifying voxels further includes the step of filtering the data to remove noise. 
     
     
         5 . The method of  claim 1 , wherein the first material decomposition method at step f) includes using volume constrained non-negative linear least squares to aid material identification. 
     
     
         6 . The computer implemented method of  claim 1 , wherein the second material decomposition method applied at step f) includes the further steps of:
 a). defining excluded material combinations in the data set pre-determined to be disallowed;   b). calculating, for each of a plurality of reconstructed voxels contained within the data set, a sparse solution for a maximum number of material combinations using a least squares (12 norm) minimisation subject to a 0-norm constraint inequality for enforcing a maximum number of materials per voxel by combinatorial search over the not excluded material combinations during the least squares minimisation using a processor;   c). applying rejection criteria to the sparse solutions of c) while performing the combinatorial search to identify one or more materials, wherein the step of applying rejection criteria to the sparse solutions includes the step of setting acceptable solution ranges for each material, and rejecting any sub-problem with a solution outside these ranges.   
     
     
         7 . The method of  claim 1 , wherein the image data is medical data and the low density voxels represent soft tissue. 
     
     
         8 . The method of  claim 1 , wherein the image data is medical data and the high density voxels represent bone, implants, metals and/or contrasting agents. 
     
     
         9 . A system for selecting and presenting a material decomposition algorithm using image data generated using a multi-energy computer tomography (CT) imaging system using three or more energy bands, the system comprising:
 a) a receiver for receiving a reconstructed or non-reconstructed multi-energy CT image data set produced using three or more energy bands in relation to an object;   b) a data storage medium for storing said data;   c) a processor for processing the data using one or more reconstruction techniques to produce a data set of reconstructed voxels;   d) a processor for comparing each voxel in the reconstructed data set to a reference set of material signal amplitude and noise properties stored on a data storage medium;   e) a processor for classifying the voxels into air, low density or high density voxels;   f) a processor for selecting a first material decomposition method to be applied to low density voxels and/or selecting a second material decomposition method to be applied to higher density voxels; and   g) an interface for presenting the selected first and/or second decomposition method to a user.   
     
     
         10 . The system of  claim 9 , wherein the system includes a multi-energy CT scanning system using three of more energy bands for scanning an object using a multi- to produce an image data set and a transmitter for sending image data to a receiver. 
     
     
         11 . A computer-implemented method for selecting a material decomposition algorithm using image data generated using image data generated using a multi-energy computer tomography (CT) imaging system using three or more energy bands, using a computer program product, the method including;
 a) scanning an object using a multi-energy CT system using three or more energy bands to produce an image data set;   b) storing said data set on a data storage medium;   c) processing the data using one or more reconstruction techniques to produce a data set of reconstructed voxels;   d) comparing each voxel in the reconstructed data set to a reference set of material signal amplitude and noise properties stored on a data storage medium;   e) classifying the voxels into air, low density or high density voxels;   f) based on the determination of e) selecting a first material decomposition method to be applied to low density voxels and/or selecting a second material decomposition method to be applied to higher density voxels; and   g) presenting the selected first and/or second decomposition method to an interface.   
     
     
         12 . The method of  claim 11 , wherein the image data is medical data and the low density voxels represent soft tissue. 
     
     
         13 . The method of  claim 11 , wherein the image data is medical data and the high density voxels represent bone, implants, metals and/or contrasting agents.