Optimized implementation of back projection for computed tomography (CT)

1. A computer-implemented method for down-sampling and restoring perspective geometry data and a weight table associated with a computed tomography device, the method comprising:

generating perspective geometry data associated with the computed tomography device;

generating a weight table associated with the computed tomography device;

down-sampling the weight table and the perspective geometry data to reduce an amount of data to be stored;

storing the down-sampled perspective geometry data and the down-sampled weight table associated with the computed tomography device in a storage device; and

reconstructing, via interpolation, the perspective geometry data and the weight table from the down-sampled perspective geometry and the down-sampled weight table, respectively, using a computing processing unit during a tomographic volume reconstruction process for an object scanned using the computed tomography device.

2. The method of claim 1 , wherein the computing processing unit is a graphical processing unit (GPU).

3. The method of claim 2 , wherein the down-sampled weight table is stored as a texture surface by the GPU.

4. The method of claim 1 further comprising:

generating a three-dimensional tomographic volume of the object using the perspective geometry data and the weight table during the tomographic volume reconstruction process.

5. The method of claim 4 , wherein the three-dimensional tomographic volume is generated using a back-projection technique, algebraic reconstruction technique (ART) or maximum likelihood expectation maximization (MLEM) technique for the computed tomography.

6. The method of claim 1 , wherein the weight table is a two-dimensional table.

7. The method of claim 1 , wherein the weight table includes a weight for each voxel within a geometry of the object being scanned for computed tomography.

8. The method of claim 7 , wherein the weight for a voxel is equal to the ratio of a distance between the voxel and an energy source used for the computed tomography to a distance between the voxel and a detector used for the computed tomography.

9. The method of claim 1 , wherein the down-sampled weight table is calculated from the weigh table using one or more of bilinear down-sampling, bicubic down-sampling or down-sampling in the frequency domain.

10. A system comprising:

a processor for:

generating perspective geometry associated with a computed tomography device;

generating a weight table associated with the computed tomography device;

down-sampling the weight table and the perspective geometry data to reduce an amount of data to be stored;

storing the down-sampled perspective geometry data and the down-sampled weight table associated with the computed tomography device in a storage device; and

reconstructing, via interpolation, the perspective geometry data and the weight table from the down-sampled perspective geometry and the down-sampled weight table, respectively during a tomographic volume reconstruction process for an object scanned using the computed tomography device.

11. The system of claim 10 , wherein the processor is a graphical processing unit (GPU).

12. The system of claim 11 , wherein the down-sampled weight table is stored as a texture surface by the GPU.

13. The system of claim 10 , wherein the processor further generates a three-dimensional tomographic volume of the object using the perspective geometry data and the weight table during the tomographic volume reconstruction process.

14. The system of claim 13 , wherein the three-dimensional tomographic volume is generated using a back-projection technique, algebraic reconstruction technique (ART) or maximum likelihood expectation maximization (MLEM) technique for the computed tomography.

15. The system of claim 10 , wherein the weight table is a two-dimensional table.

16. The system of claim 10 , wherein the weight table includes a weight for each voxel within a geometry of the object being scanned for computed tomography.

17. The system of claim 16 , wherein the weight for a voxel is equal to the ratio of a distance between the voxel and an energy source used for the computed tomography to a distance between the voxel and a detector used for the computed tomography.

18. The system of claim 10 , wherein the down-sampled weight table is calculated from the weigh table using one or more of bilinear down-sampling, bicubic down-sampling or down-sampling in the frequency domain.

19. A non-transitory computer-readable medium for storing instructions that, when executed on a computer, down-sample and restore perspective geometry data and a weight table associated with a computed tomography device, the medium storing one or more instructions for:

generating perspective geometry data associated with the computed tomography device;

generating a weight table associated with the computed tomography device;

down-sampling the weight table and the perspective geometry data to reduce an amount of data to be stored;

storing the down-sampled perspective geometry data and the down-sampled weight table associated with the computed tomography device in a storage device; and

reconstructing, via interpolation, the perspective geometry data and the weight table from the down-sampled perspective geometry and the down-sampled weight table, respectively, using a graphical processing unit (GPU) during a tomographic volume reconstruction process for an object scanned using the computed tomography device, wherein the down-sampled weight table is stored as a texture surface by the GPU.

20. The non-transitory medium of claim 19 further storing one or more instructions for:

generating a three-dimensional tomographic volume of the object using the perspective geometry data and the weight table during the tomographic volume reconstruction process.