An Effective CUDA Parallelization of Projection in Iterative Tomography Reconstruction

Lizhe Xie; Yining Hu; Bin Yan; Lin Wang; Benqiang Yang; Wenyuan Liu; Libo Zhang; Limin Luo; Huazhong Shu; Yang Chen

doi:10.1371/journal.pone.0142184

An Effective CUDA Parallelization of Projection in Iterative Tomography Reconstruction

PLoS One. 2015 Nov 30;10(11):e0142184. doi: 10.1371/journal.pone.0142184. eCollection 2015.

Authors

Lizhe Xie¹, Yining Hu^{2

3

4}, Bin Yan¹, Lin Wang¹, Benqiang Yang⁵, Wenyuan Liu⁵, Libo Zhang⁵, Limin Luo^{2

3

4}, Huazhong Shu^{2

3

4}, Yang Chen^{2

3

4}

Affiliations

¹ Oral Hospital of Jiangsu Province, Affiliated to Nanjing Medical University, Jiangsu, China.
² Centre de Recherche en Information Biomedicale Sino-Francais (LIA CRIBs), Rennes, France.
³ The Laboratory of Image Science and Technology, Southeast University, Nanjing, China.
⁴ The Key Laboratory of Computer Network and Information Integration (Southeast University), Ministry of Education, Beijing, China.
⁵ Department of Radiology, General Hospital of Shenyang Military Area Command, Shenhe District, Shenyang, 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.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Algorithms*
Image Processing, Computer-Assisted / methods*
Tomography, X-Ray Computed / methods*

Grants and funding

This research was supported by National Natural Science Foundation under grants (81370040, 81530060, 61405033, 31100713). This work was also supported by the grant of Natural Science Foundation of Jiangsu Province under grant BK20130629, and the Qing Lan Project in Jiangsu Province.