Object Identification in Binary Tomographic Images Using GPGPUs

Research output: Contribution to journalArticle

Abstract

The authors present a hybrid OpenCL CPU/GPU algorithm for identification of connected structures inside black and white 3D scientific data. This algorithm exploits parallelism both at CPU and GPGPU levels, but the work is predominantly done in GPUs. The underlying context of this work is the structural characterization of composite materials via tomography. The algorithm allows us to later infer location and morphology of objects inside composite materials. Moreover, execution times are very low thus allowing us to process large data sets, but within acceptable running times. Intermediate solutions are computed independently over a partition of the spatial domain, following the data parallelism paradigm, and then integrated both at GPU and CPU levels, using parallel multi-cores. The authors consistently explore parallelism both at the CPU level, by allowing the CPU stage to run in multiple concurrent threads, and at the GPU level with massive parallelism and concurrent data transfers and kernel executions.
Original languageEnglish
Pages (from-to)40-56
JournalInternational Journal of Creative Interfaces and Computer Graphics
Volume4
Issue number2
DOIs
Publication statusPublished - 2013

Fingerprint

Binary images
Program processors
Composite materials
Data transfer
Tomography
Graphics processing unit

Cite this

@article{5f48da15cad84a10bcba808cb5c55b24,
title = "Object Identification in Binary Tomographic Images Using GPGPUs",
abstract = "The authors present a hybrid OpenCL CPU/GPU algorithm for identification of connected structures inside black and white 3D scientific data. This algorithm exploits parallelism both at CPU and GPGPU levels, but the work is predominantly done in GPUs. The underlying context of this work is the structural characterization of composite materials via tomography. The algorithm allows us to later infer location and morphology of objects inside composite materials. Moreover, execution times are very low thus allowing us to process large data sets, but within acceptable running times. Intermediate solutions are computed independently over a partition of the spatial domain, following the data parallelism paradigm, and then integrated both at GPU and CPU levels, using parallel multi-cores. The authors consistently explore parallelism both at the CPU level, by allowing the CPU stage to run in multiple concurrent threads, and at the GPU level with massive parallelism and concurrent data transfers and kernel executions.",
keywords = "Parallelization, OpenCL, CPU, 3D Image Data Processing, GPGPU, Tomography",
author = "Bruno Preto and Birra, {Fernando Pedro Reino da Silva} and Lopes, {Adriano Martins} and Medeiros, {Pedro Ab{\'i}lio Duarte de}",
year = "2013",
doi = "10.4018/ijcicg.2013070103",
language = "English",
volume = "4",
pages = "40--56",
journal = "International Journal of Creative Interfaces and Computer Graphics",
number = "2",

}

TY - JOUR

T1 - Object Identification in Binary Tomographic Images Using GPGPUs

AU - Preto, Bruno

AU - Birra, Fernando Pedro Reino da Silva

AU - Lopes, Adriano Martins

AU - Medeiros, Pedro Abílio Duarte de

PY - 2013

Y1 - 2013

N2 - The authors present a hybrid OpenCL CPU/GPU algorithm for identification of connected structures inside black and white 3D scientific data. This algorithm exploits parallelism both at CPU and GPGPU levels, but the work is predominantly done in GPUs. The underlying context of this work is the structural characterization of composite materials via tomography. The algorithm allows us to later infer location and morphology of objects inside composite materials. Moreover, execution times are very low thus allowing us to process large data sets, but within acceptable running times. Intermediate solutions are computed independently over a partition of the spatial domain, following the data parallelism paradigm, and then integrated both at GPU and CPU levels, using parallel multi-cores. The authors consistently explore parallelism both at the CPU level, by allowing the CPU stage to run in multiple concurrent threads, and at the GPU level with massive parallelism and concurrent data transfers and kernel executions.

AB - The authors present a hybrid OpenCL CPU/GPU algorithm for identification of connected structures inside black and white 3D scientific data. This algorithm exploits parallelism both at CPU and GPGPU levels, but the work is predominantly done in GPUs. The underlying context of this work is the structural characterization of composite materials via tomography. The algorithm allows us to later infer location and morphology of objects inside composite materials. Moreover, execution times are very low thus allowing us to process large data sets, but within acceptable running times. Intermediate solutions are computed independently over a partition of the spatial domain, following the data parallelism paradigm, and then integrated both at GPU and CPU levels, using parallel multi-cores. The authors consistently explore parallelism both at the CPU level, by allowing the CPU stage to run in multiple concurrent threads, and at the GPU level with massive parallelism and concurrent data transfers and kernel executions.

KW - Parallelization

KW - OpenCL

KW - CPU

KW - 3D Image Data Processing

KW - GPGPU

KW - Tomography

U2 - 10.4018/ijcicg.2013070103

DO - 10.4018/ijcicg.2013070103

M3 - Article

VL - 4

SP - 40

EP - 56

JO - International Journal of Creative Interfaces and Computer Graphics

JF - International Journal of Creative Interfaces and Computer Graphics

IS - 2

ER -