Sparse MRI data collection and classification using machine learning

1. A method for implementing a sparse sampling strategy for acquiring MRI data, comprising:

collecting and labeling a training dataset of MRI scans for a predetermined diagnostic;

selecting a sampling shape and associated parameter values wherein the sampling shape is a diamond-like shape;

sampling each MM scan in the training data set using the sampling shape and associated parameter values to generate a set of sparse samples;

training a neural network using the sparse samples and assigning an accuracy to a resulting trained neural network; and

adjusting the associated parameter values, and repeating the sampling and training until optimized parameter values are established.

2. The method of claim 1 , further comprising:

configuring an MRI machine with a sparse sampling strategy that includes the sampling shape and optimized parameter values.

3. The method of claim 2 , further comprising:

acquiring an MM scan from a patient using the sparse sampling strategy; and

utilizing the trained neural network to generate a classification of the MM scan.

4. The method of claim 3 , wherein the classification provides a diagnosis and a confidence level.

5. The method of claim 4 , wherein the trained neural network includes a deep learning model that provides a severity of the diagnosis.

6. The method of claim 1 , wherein the training dataset includes raw k-space data and original acquisition parameters.

7. A system for implementing a sparse sampling strategy for acquiring MRI data, comprising:

a system for collecting and labeling a training dataset of MRI scans for a predetermined diagnostic;

a system for selecting a sampling shape and associated parameter values wherein the sampling shape is a diamond-like shape;

a sampling system for sampling each MRI scan in the training data set using the sampling shape and associated parameter values to generate a set of sparse samples;

a training system for training a neural network using the sparse samples and assigning an accuracy to the neural network; and

a parameter optimization system that adjusts the associated parameter values and repeatedly runs the sampling system and training system until optimized parameter values are established.

8. The system of claim 7 , further comprising:

an Mill machine configured with a sparse sampling strategy that includes the sampling shape and optimized parameter values.

9. The system of claim 7 , further comprising:

a classification system that utilizes the trained neural network to classify an Mill scan obtained from an MM machine configured with a sparse sampling strategy that includes the sampling shape and optimized parameter values.

10. The system of claim 9 , wherein the classification system provides a diagnosis and a confidence level.

11. The system of claim 10 , wherein the trained neural network includes a deep learning model that provides a severity of the diagnosis.

12. The system of claim 7 , wherein the training dataset includes raw k-space data and original acquisition parameters.

13. A computer program product stored on a non-transitory computer readable storage medium, which when executed by a computing system, provides a sparse sampling strategy for acquiring MRI data, the program product comprising:

program code that collects labels a training dataset of MRI scans for a predetermined diagnostic;

program code for selecting a sampling shape and associated parameter values wherein the sampling shape is a diamond-like shape;

program code for sampling each MRI scan in the training data set using the sampling shape and associated parameter values to generate a set of sparse samples;

program code for training a neural network using the sparse samples and assigning an accuracy to a resulting trained neural network; and

program code for adjusting the associated parameter values, and repeating the sampling and training until optimized parameter values are established.

14. The program product of claim 13 , further comprising:

program code for inputting an acquired MRI scan from a patient using the sparse sampling strategy; and

program code for utilizing the trained neural network to generate a classification of the MRI scan.

15. The program product of claim 14 , wherein the classification provides a diagnosis and a confidence level.

16. The program product of claim 13 , wherein the trained neural network includes a deep learning model that provides a severity of the diagnosis.

17. The program product of claim 13 , wherein the training dataset includes raw k-space data and original acquisition parameters.