Limits...
An Effective CUDA Parallelization of Projection in Iterative Tomography Reconstruction.

Xie L, Hu Y, Yan B, Wang L, Yang B, Liu W, Zhang L, Luo L, Shu H, Chen Y - PLoS ONE (2015)

Bottom Line: In this paper, a strategy of Fixed Sampling Number Projection (FSNP) is proposed to ensure the operation synchronization in the ray-driven projection with Graphical Processing Unit (GPU).We validate the performance of this FSNP approach using both simulated and real cone-beam CT data.Experimental results show that compare to the conventional approach, the proposed FSNP method together with texture fetching is 10~16 times faster than the conventional approach based on global memory, and thus leads to more efficient iterative algorithm in CT reconstruction.

View Article: PubMed Central - PubMed

Affiliation: Oral Hospital of Jiangsu Province, Affiliated to Nanjing Medical University, Jiangsu, China.

ABSTRACT
Projection and back-projection are the most computationally intensive parts in Computed Tomography (CT) reconstruction, and are essential to acceleration of CT reconstruction algorithms. Compared to back-projection, parallelization efficiency in projection is highly limited by racing condition and thread unsynchronization. In this paper, a strategy of Fixed Sampling Number Projection (FSNP) is proposed to ensure the operation synchronization in the ray-driven projection with Graphical Processing Unit (GPU). Texture fetching is also used utilized to further accelerate the interpolations in both projection and back-projection. We validate the performance of this FSNP approach using both simulated and real cone-beam CT data. Experimental results show that compare to the conventional approach, the proposed FSNP method together with texture fetching is 10~16 times faster than the conventional approach based on global memory, and thus leads to more efficient iterative algorithm in CT reconstruction.

Show MeSH

Related in: MedlinePlus

One 2D slice reconstructed from one rat data scanned from a micro CT.(A): FDK with Ram-Lak Filter; (B): OSEM with linear interpolation based FSNP; (C): OSEM with blob based FSNP. (D), (E) and (F) are the zoomed regions from (A), (B) and (C), respectively.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4664243&req=5

pone.0142184.g007: One 2D slice reconstructed from one rat data scanned from a micro CT.(A): FDK with Ram-Lak Filter; (B): OSEM with linear interpolation based FSNP; (C): OSEM with blob based FSNP. (D), (E) and (F) are the zoomed regions from (A), (B) and (C), respectively.

Mentions: The proposed projection methods are also validated by real scan data from a micro CT system. A rat was scanned under 40kV tube voltage and 200mA tube current. The projection sequence contains 360 projections over 360°. The projection image size is 922×748 with pixel size 0.1mm2; the reconstruction volume size is 512×512×512 with voxel size 0.085mm2, SOD = 83.65mm, SDD = 167.3mm. The FDK algorithm with Ram-Lak filter and the OSEM algorithm with 30 subsets and 100 iterations were performed. All the methods listed are with voxel-driven back-projector. Different from the observation in Fig 5, the reconstruction results Fig 7 indicates that, compared to the linear interpolation based FSNP, the blob based FSNP provides results with better edge preservation (see the arrows). Nevertheless, we can also see that the improved edge preservation is also accompanied with amplified noise in the reconstructed images.


An Effective CUDA Parallelization of Projection in Iterative Tomography Reconstruction.

Xie L, Hu Y, Yan B, Wang L, Yang B, Liu W, Zhang L, Luo L, Shu H, Chen Y - PLoS ONE (2015)

One 2D slice reconstructed from one rat data scanned from a micro CT.(A): FDK with Ram-Lak Filter; (B): OSEM with linear interpolation based FSNP; (C): OSEM with blob based FSNP. (D), (E) and (F) are the zoomed regions from (A), (B) and (C), respectively.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4664243&req=5

pone.0142184.g007: One 2D slice reconstructed from one rat data scanned from a micro CT.(A): FDK with Ram-Lak Filter; (B): OSEM with linear interpolation based FSNP; (C): OSEM with blob based FSNP. (D), (E) and (F) are the zoomed regions from (A), (B) and (C), respectively.
Mentions: The proposed projection methods are also validated by real scan data from a micro CT system. A rat was scanned under 40kV tube voltage and 200mA tube current. The projection sequence contains 360 projections over 360°. The projection image size is 922×748 with pixel size 0.1mm2; the reconstruction volume size is 512×512×512 with voxel size 0.085mm2, SOD = 83.65mm, SDD = 167.3mm. The FDK algorithm with Ram-Lak filter and the OSEM algorithm with 30 subsets and 100 iterations were performed. All the methods listed are with voxel-driven back-projector. Different from the observation in Fig 5, the reconstruction results Fig 7 indicates that, compared to the linear interpolation based FSNP, the blob based FSNP provides results with better edge preservation (see the arrows). Nevertheless, we can also see that the improved edge preservation is also accompanied with amplified noise in the reconstructed images.

Bottom Line: In this paper, a strategy of Fixed Sampling Number Projection (FSNP) is proposed to ensure the operation synchronization in the ray-driven projection with Graphical Processing Unit (GPU).We validate the performance of this FSNP approach using both simulated and real cone-beam CT data.Experimental results show that compare to the conventional approach, the proposed FSNP method together with texture fetching is 10~16 times faster than the conventional approach based on global memory, and thus leads to more efficient iterative algorithm in CT reconstruction.

View Article: PubMed Central - PubMed

Affiliation: Oral Hospital of Jiangsu Province, Affiliated to Nanjing Medical University, Jiangsu, China.

ABSTRACT
Projection and back-projection are the most computationally intensive parts in Computed Tomography (CT) reconstruction, and are essential to acceleration of CT reconstruction algorithms. Compared to back-projection, parallelization efficiency in projection is highly limited by racing condition and thread unsynchronization. In this paper, a strategy of Fixed Sampling Number Projection (FSNP) is proposed to ensure the operation synchronization in the ray-driven projection with Graphical Processing Unit (GPU). Texture fetching is also used utilized to further accelerate the interpolations in both projection and back-projection. We validate the performance of this FSNP approach using both simulated and real cone-beam CT data. Experimental results show that compare to the conventional approach, the proposed FSNP method together with texture fetching is 10~16 times faster than the conventional approach based on global memory, and thus leads to more efficient iterative algorithm in CT reconstruction.

Show MeSH
Related in: MedlinePlus