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

Non-uniform distributed samples points in projection geometry.(a) cubic FOV, (b) cylinder FOV, (c) sphere FOV
© Copyright Policy
Related In: Results  -  Collection

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

pone.0142184.g002: Non-uniform distributed samples points in projection geometry.(a) cubic FOV, (b) cylinder FOV, (c) sphere FOV

Mentions: The sampling point numbers along different projection lines are not identical due to the varying intersection lengths for different projection lines. The varying sampling numbers will lead to unsynchronized operations in threads or blocks for PRPT or PRPB modes, which result in lowered parallelization efficiency. In this paper, a method of Fixed Sampling Number Projection (FSNP) is proposed to overcome this. As illustrated by Fig 2, this FSNP approach assumes that the X-ray attenuation is negligible outside a pre-specified field of view (FOV). We fix the sampling number along each ray and perform uniform sampling on the line segments inside the FOV, thus allows an easy synchronization for both the PRPT and PRPB modes.


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)

Non-uniform distributed samples points in projection geometry.(a) cubic FOV, (b) cylinder FOV, (c) sphere FOV
© Copyright Policy
Related In: Results  -  Collection

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

pone.0142184.g002: Non-uniform distributed samples points in projection geometry.(a) cubic FOV, (b) cylinder FOV, (c) sphere FOV
Mentions: The sampling point numbers along different projection lines are not identical due to the varying intersection lengths for different projection lines. The varying sampling numbers will lead to unsynchronized operations in threads or blocks for PRPT or PRPB modes, which result in lowered parallelization efficiency. In this paper, a method of Fixed Sampling Number Projection (FSNP) is proposed to overcome this. As illustrated by Fig 2, this FSNP approach assumes that the X-ray attenuation is negligible outside a pre-specified field of view (FOV). We fix the sampling number along each ray and perform uniform sampling on the line segments inside the FOV, thus allows an easy synchronization for both the PRPT and PRPB modes.

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