Three-dimensional forward and back projection methods
Methods provided for forward and back-projection, which are referred to as separable footprint (SF) projectors: exemplified by the SF-TR and SF-TT projectors. These methods approximate the voxel footprint functions as 2D separable functions. Because of the separability of these footprint functions, calculating their integrals over a detector cell is greatly simplified and can be implemented efficiently. In some embodiments, the SF-TR projector uses trapezoid functions in the transaxial direction and rectangular functions in the axial direction. In some embodiments, the SF-TT projector uses trapezoid functions in both the axial and transaxial directions. Simulations and experiments showed that both SF projector methods are more accurate than conventional distance-driven (DD) projectors. Moreover, the SF-TT projector is more accurate than the SF-TR projector for rays associated with large cone angles. In some embodiments, the SF-TR projector has similar computation speed with the DD projector and the SF-TT projector is about two times slower.
1. A processor-implemented method for forward and back projection for image construction, the method comprising:
computing projected coordinates of a plurality of 3D voxels on a detector surface;
approximating 2D footprints of said plurality of voxels as separable functions, said separable functions being a computational product of a 1D axial footprint function and a 1D transaxial footprint function;
approximating 2D detector blur as a computational product of a 1D axial blur function and a 1D transaxial blur function; and
determining the contribution of said plurality of voxels to detector cells or vice versa using computational sums and products of said 1D axial footprint function, said 1D transaxial footprint function, said 1D axial blur function, and said 1D transaxial blur function in accordance with said projected coordinates.
2. The method according to claim 1 ,
wherein said 1D transaxial footprint function comprises trapezoidal functions and said 1D axial footprint function comprises rectangular functions.
3. The method according to claim 1
wherein said 1D transaxial footprint function and said 1D axial footprint function each comprises trapezoidal functions.
4. The method according to claim 1
wherein said 1D transaxial footprint function and said 1D axial footprint function each comprises a combination of trapezoidal functions and rectangular functions.
5. The method according to claim 1 , further comprising:
determining vertices of said 1D axial footprint function and said 1D transaxial footprint function using said projected coordinates of said plurality of 3D voxels.
6. The method according to claim 1 wherein parameters of the step of computing projected coordinates of said plurality of voxels comprises connecting a source and computing said projected coordinates in response to said boundaries of said plurality of voxels and said detector surface.
7. The method according to claim 1 wherein parameters of the step of determining the contribution of said plurality of voxels to detector cells or vice versa comprises:
computing 2D blurred footprints of said plurality of 3D voxels as computational products of 1D axial blurred footprints and 1D transaxial blurred footprints, said 1D axial blurred footprints being 1D convolutions of said 1D axial footprint functions and said 1D axial blur functions, said 1D transaxial blurred footprints being 1D convolutions of said 1D transaxial footprint functions and said 1D transaxial blur functions.
8. The method according to claim 1 wherein said determining the contribution of said plurality of voxels to said detector cells or vice versa comprises at least one of a forward projection and a back projection operation followed or preceded by detector-cell-dependent amplitude scaling factors.
9. The method according to claim 1 wherein the determining the contribution of said plurality of voxels to said detector cells or vice versa comprises at least one of a forward projection and a back projection operation followed or preceded by voxel-dependent scaling factors.
10. The method according to claim 1 wherein the determining the contribution of said plurality of voxels to said detector cells or vice versa comprises at least one of a forward projection and a back projection operation followed or preceded by a combination of detector-cell-dependent and voxel-dependent scaling factors.
11. A projector or back-projector system having a non-transitory computer readable medium encoded with an executable program for image reconstruction, said projector or back-projector system comprising:
a computational system using computational operations occurring separately in axial and transaxial directions that exploit the separability of approximation of a footprint function, said computational operations comprising:
computing projected coordinates of a plurality of 3D voxels on a detector surface;
approximating 2D footprints of said plurality of voxels as separable functions, said separable functions being a computational product of a 1D axial footprint function and a 1D transaxial footprint function;
approximating 2D detector blur as a computational product of a 1D axial blur function and a 1D transaxial blur function; and
determining the contribution of said plurality of voxels to detector cells or vice versa using computational sums and products of said 1D axial footprint function, said 1D transaxial footprint function, said 1D axial blur function, and said 1D transaxial blur function in accordance with said projected coordinates.
12. The system according to claim 11 wherein said computational operations occur separately and in parallel in axial and transaxial directions.
13. The system according to claim 11 wherein said 1D transaxial footprint function comprises trapezoidal functions and said 1D axial footprint function comprises rectangular functions.
14. The system according to claim 11 wherein said 1D transaxial footprint function and said 1D axial footprint function each comprises trapezoidal functions.
15. The system according to claim 11 wherein said 1D transaxial footprint function and said 1D axial footprint function each comprises rectangular functions.
16. The system according to claim 11 wherein said 1D transaxial footprint function and said 1D axial footprint function each comprises a combination of trapezoidal functions and rectangular functions.
17. The method according to claim 1 wherein said 1D transaxial footprint function and said 1D axial footprint function each comprises rectangular functions.
18. The method according to claim 1 wherein said 1D axial blur function and said 1D transaxial blur function each comprises rectangular functions.