Single- and multi-modality alignment of medical images in the presence of non-rigid deformations using phase correlation

1. A method comprising:

receiving, from at least one of an MRI imaging device or a CT imaging device, a first medical image of a first registered volume of a patient taken at a first time and at a first position and a second medical image of a second registered volume of the patient taken at a second time and at a second position;

comparing the first medical image and the second medical image using a phase correlation method, the comparing comprising:

identifying a peak in a normalized cross-power spectrum of the first registered volume and the second registered volume;

finding a maximum intensity of a Fourier transform of the normalized cross-power spectrum;

selecting, from a plurality of voxels having intensities greater than a threshold, a voxel for which a sum of voxel intensities of neighboring voxels around the voxel is highest; and

calculating, based at least on the voxel, at least one of a translation and a rotation required to properly align the first and second medical images in a common registration grid;

determining, based on the calculating, a change to at least one of a physical location and a physical orientation of the patient; and

correcting, based on the determined change, the second position of the patient to more closely conform to the first position of the patient.

2. A method as in claim 1 , further comprising downsampling the first registered volume and the second registered volume to create the common registration grid having a lower resolution than either of the first registered volume and the second registered volume.

3. A method as in claim 1 , further comprising:

determining that the first registered volume and the second registered volume comprise different resolutions; and

resampling the first registered volume and/or the second registered volume on the common registration grid.

4. A method as in claim 3 , wherein a common resolution along each dimension of the common registration grid is set to a coarser of a first initial resolution of the first registered volume and a second initial resolution of the second registered volume.

5. A method as in claim 1 , wherein the neighboring voxels are in a window defined as a fraction of a number of voxels along each dimension of the common registration grid.

6. A method as in claim 1 , further comprising refining a position of the peak by calculating a centroid of the voxel intensities of the neighboring voxels.

7. A method as in claim 1 , further comprising:

applying the change to the physical location and/or the physical orientation of the patient; and

performing a medical procedure on the patient after applying the change.

8. A method as in claim 7 , wherein the medical procedure comprises at least one of a radiation treatment and a surgical procedure.

9. A method as in claim 1 , wherein the first image and the second image are obtained using a same imaging modality.

10. A method as in claim 1 , wherein the first image and the second image are obtained using different imaging modalities.

11. A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:

receiving, from at least one of an MRI imaging device or a CT imaging device, a first medical image of a first registered volume of a patient taken at a first time and a first position and a second medical image of a second registered volume of the patient taken at a second time and a second position;

comparing the first medical image and the second medical image using a phase correlation method, the comparing comprising:

identifying a peak in a normalized cross-power spectrum of the first registered volume and the second registered volume;

finding a maximum intensity of a Fourier transform of the normalized cross-power spectrum;

selecting, from a plurality of voxels having intensities greater than a threshold, a voxel for which a sum of voxel intensities of neighboring voxels around the voxel is highest; and

calculating, based at least on the voxel, at least one of a translation and a rotation required to properly align the first and second medical images in a common registration grid;

determining, based on the calculating, a change to at least one of a physical location and a physical orientation of the patient; and

correcting, based on the determined change, the second position of the patient to more closely conform to the first position of the patient.

12. A computer program product as in claim 11 , wherein the operations further comprise downsampling the first registered volume and the second registered volume to create the common registration grid having a lower resolution than either of the first registered volume and the second registered volume.

13. A computer program product as in claim 11 , wherein the operations further comprise:

determining that the first registered volume and the second registered volume comprise different resolutions; and

resampling the first registered volume and/or the second registered volume on the common registration grid.

14. A system comprising:

computer hardware configured to perform operations comprising:

receiving, from at least one of an MRI imaging device or a CT imaging device, a first medical image of a first registered volume of a patient taken at a first time and a first position and a second medical image of a second registered volume of the patient taken at a second time and a second position;

comparing the first medical image and the second medical image using a phase correlation method, the comparing comprising:

identifying a peak in a normalized cross-power spectrum of the first registered volume and the second registered volume;

finding a maximum intensity of a Fourier transform of the normalized cross-power spectrum;

selecting, from a plurality of voxels having intensities greater than a threshold, a voxel for which a sum of voxel intensities of neighboring voxels around the voxel is highest; and

calculating, based at least on the voxel, at least one of a translation and a rotation required to properly align the first and second medical images in a common registration grid;

determining, based on the calculating, a change to at least one of a physical location and a physical orientation of the patient; and

correcting, based on the determined change, the second position of the patient to more closely conform to the first position of the patient.

15. A system as in claim 14 , further comprising at least one imaging device for generating the first and second medical images.

16. A system as in claim 14 , the calculating further comprising performing a three-dimensional Fourier transform on at least one of the first medical image or the second medical image, wherein the first medical image is a three-dimensional image of the first registered volume of the patient and the second medical image is a three-dimensional image of the second registered volume of the patient.

17. A method as in claim 1 , the phase correlation method comprising generating a cross-power spectrum of the first medical image and the second medical image and the calculating of the at least one of the translation and the rotation being based at least on the cross-power spectrum.