System and method for image processing with highly undersampled imaging data
A system and method for processing highly undersampled multi-echo spin-echo data by linearizing the slice-resolved extended phase graph model generates highly accurate T 2 maps with indirect echo compensation. Principal components are used to linearize the signal model to estimate the T 2 decay curves which can be fitted to the slice-resolved model for T2 estimation. In another example of image processing for highly undersampled data, a joint bi-exponential fitting process can compensate for image variations within a voxel and thus provide partial voxel compensation to produce more accurate T 2 maps.
1. A system comprising:
a magnetic resonance imaging (MRI) device configured to generate imaging data, the generated image data being generated using predetermined pulse sequence and being highly undersampled;
an interface configured to receive the imaging data generated by the MRI device; and
a processor configured to process the received imaging data using a linear approximation to a signal model characterizing a T 2 decay to thereby generate a corrected T 2 estimation map to compensate for errors in the highly undersampled generated imaging data caused by imperfections in the radio frequency refocusing pulses associated with the MRI device or multiple components due to the presence of different tissue species within a voxel to thereby generate and display a more accurate MRI image.
2. The system of claim 1 wherein the generated imaging data is radial fast-spin-echo acquisition data.
3. The system of claim 1 wherein the processor is configured to process the received imaging data to compensate for indirect echoes in the generated image data.
4. The system of claim 1 wherein the processor is configured to process the received imaging data to generate an estimation of the T 2 relaxation time.
5. The system of claim 1 wherein the processor is configured to process the received imaging data to apply a linearization model to the received image data and to compensate for indirect echoes in the generated image data.
6. The system of claim 1 , further comprising:
a set of training curves generated for a range of expected T 2 values; and
a set of coefficients derived from a linear approximation model and the set of training curves, the coefficients being representative of T 2 decay curves in the presence of indirect echoes,
wherein the processor is further configured to apply the set of coefficients to the imaging data to compensate for imperfections in the refocusing pulses in the MRI device that distort the generated imaging data to thereby generate the T 2 map.
7. The system of claim 6 wherein the coefficients are derived from a group comprising principal components, a manifold, and a dictionary.
8. The system of claim 1 wherein the processor is configured to process the received imaging data from an imaging sample containing more than one tissue or chemical component within a voxel.
9. The system of claim 8 wherein the processor is configured to process the received imaging data using a set of initial lesion intensity values and a set of initial background intensity values for each voxel in a designated region of interest (ROI) containing an image of a lesion and to analyze the sets of initial intensity values in a fitting process to thereby generate the T 2 map representative of the lesion.
10. The system of claim 9 wherein the processor is configured to analyze the sets of initial intensity values using a joint bi-exponential fitting algorithm.
11. A method for processing imaging data generated by a magnetic resonance imaging (MRI) data comprising:
receiving the imaging data, the imaging data being generated using a predetermined pulse sequence and being highly undersampled; and
processing the received imaging data using a linear approximation to the signal model to thereby generate a corrected T 2 estimation map to compensate for errors in the highly undersampled generated imaging data caused by indirect echoes generated by imperfections in the radio frequency refocusing pulses associated with the MRI device or multiple components due to the presence of different tissue species within a voxel to thereby generate and display a more accurate MRI image.
12. The method of claim 11 wherein processing the received imaging data comprises generating a T 2 estimation map to thereby compensate for indirect echoes in the imaging data.
13. The method of claim 11 wherein processing the received imaging data comprises generating an estimation of the T 2 relaxation time.
14. The method of claim 11 wherein processing the received imaging data comprises applying a linearization model to the received imaging data and compensating for indirect echoes in the generated image data.
15. The method of claim 11 , further comprising:
generating a set of training curves for a range of expected T 2 values;
deriving a set of coefficients from a linear model and the set of training curves, the principal components being representative of T 2 decay curves in the presence of indirect echoes; and
applying the set of coefficients to the imaging data to compensate for refocusing pulses in the MRI device that distort the generated imaging data to thereby generate the T 2 map.
16. The method of claim 15 wherein the coefficients are derived from a group comprising principal components, a manifold, and a dictionary.
17. The method of claim 11 wherein processing the received imaging data comprises processing the received imaging data from an imaging sample containing more than one tissue or chemical component within a voxel.
18. The method of claim 17 wherein processing the received imaging data uses a set of initial lesion intensity values and a set of initial background intensity values for each voxel in a designated region of interest (ROI) containing an image of a lesion and analyzing the sets of initial intensity values in a fitting process to thereby generate the T 2 map representative of the lesion.
19. The method of claim 18 wherein analyzing the sets of initial intensity values uses a joint bi-exponential fitting algorithm.
20. A system for processing magnetic resonance imaging data from an MRI device, comprising:
an interface configured to receive the imaging data generated by the MRI device, the imaging data being generated using a predetermined pulse sequence and being highly undersampled;
a set of training curves generated from imaging data for a range of expected T 2 values;
a set of principal components derived from the set of training curves, the principal components being representative of T 2 decay curves in the presence of indirect echoes generated by imperfections in the radio frequency refocusing pulses associated with the MRI device; and
a processor configured to apply the set of principal components to the imaging data to compensate for the imperfections in the refocusing pulses used in the MRI device that distort the generated imaging data to thereby generate a corrected T 2 map that reduces T 2 value dependence on the imperfection in the MRI device and to display a more accurate MRI image.