Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures

Mikhail Smelyanskiy, David R. Holmes III, Jatin Chhugani, Alan Larson, Douglas M. Carmean, Dennis Hanson, Pradeep Dubey, Kurt Augustine, Daehyun Kim, Alan Kyker, Victor W. Lee, Anthony D. Nguyen, Larry Seiler, Richard Robb

Research output: Contribution to journalArticle

47 Citations (Scopus)

Abstract

Medical volumetric imaging requires high fidelity, high performance rendering algorithms. We motivate and analyze new volumetric rendering algorithms that are suited to modern parallel processing architectures. First, we describe the three major categories of volume rendering algorithms and confirm through an imaging scientist-guided evaluation that ray-casting is the most acceptable. We describe a thread- and data-parallel implementation of ray-casting that makes it amenable to key architectural trends of three modern commodity parallel architectures: multi-core, GPU, and an upcoming many-y-core Intel R architecture code-named Larrabee. We achieve more than an order of magnitude performance improvement on a number of large 3D medical datasets. We further describe a data compression scheme that significantly reduces data-transfer overhead. This allows our approach to scale well to large numbers of Larrabee cores.

Original languageEnglish (US)
Article number5290774
Pages (from-to)1563-1570
Number of pages8
JournalIEEE Transactions on Visualization and Computer Graphics
Volume15
Issue number6
DOIs
StatePublished - Nov 2009

Fingerprint

Volume rendering
Medical imaging
Program processors
Casting
Imaging techniques
Parallel architectures
Data compression
Data transfer
Processing
Graphics processing unit

Keywords

  • GPGPU
  • Graphics Architecture
  • Many-core Computing
  • Medical Imaging
  • Parallel Processing
  • Volume Compositing

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures. / Smelyanskiy, Mikhail; Holmes III, David R.; Chhugani, Jatin; Larson, Alan; Carmean, Douglas M.; Hanson, Dennis; Dubey, Pradeep; Augustine, Kurt; Kim, Daehyun; Kyker, Alan; Lee, Victor W.; Nguyen, Anthony D.; Seiler, Larry; Robb, Richard.

In: IEEE Transactions on Visualization and Computer Graphics, Vol. 15, No. 6, 5290774, 11.2009, p. 1563-1570.

Research output: Contribution to journalArticle

Smelyanskiy, M, Holmes III, DR, Chhugani, J, Larson, A, Carmean, DM, Hanson, D, Dubey, P, Augustine, K, Kim, D, Kyker, A, Lee, VW, Nguyen, AD, Seiler, L & Robb, R 2009, 'Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures', IEEE Transactions on Visualization and Computer Graphics, vol. 15, no. 6, 5290774, pp. 1563-1570. https://doi.org/10.1109/TVCG.2009.164
Smelyanskiy, Mikhail ; Holmes III, David R. ; Chhugani, Jatin ; Larson, Alan ; Carmean, Douglas M. ; Hanson, Dennis ; Dubey, Pradeep ; Augustine, Kurt ; Kim, Daehyun ; Kyker, Alan ; Lee, Victor W. ; Nguyen, Anthony D. ; Seiler, Larry ; Robb, Richard. / Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures. In: IEEE Transactions on Visualization and Computer Graphics. 2009 ; Vol. 15, No. 6. pp. 1563-1570.
@article{99e14a9051c4467086369ffc8764e24b,
title = "Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures",
abstract = "Medical volumetric imaging requires high fidelity, high performance rendering algorithms. We motivate and analyze new volumetric rendering algorithms that are suited to modern parallel processing architectures. First, we describe the three major categories of volume rendering algorithms and confirm through an imaging scientist-guided evaluation that ray-casting is the most acceptable. We describe a thread- and data-parallel implementation of ray-casting that makes it amenable to key architectural trends of three modern commodity parallel architectures: multi-core, GPU, and an upcoming many-y-core Intel R architecture code-named Larrabee. We achieve more than an order of magnitude performance improvement on a number of large 3D medical datasets. We further describe a data compression scheme that significantly reduces data-transfer overhead. This allows our approach to scale well to large numbers of Larrabee cores.",
keywords = "GPGPU, Graphics Architecture, Many-core Computing, Medical Imaging, Parallel Processing, Volume Compositing",
author = "Mikhail Smelyanskiy and {Holmes III}, {David R.} and Jatin Chhugani and Alan Larson and Carmean, {Douglas M.} and Dennis Hanson and Pradeep Dubey and Kurt Augustine and Daehyun Kim and Alan Kyker and Lee, {Victor W.} and Nguyen, {Anthony D.} and Larry Seiler and Richard Robb",
year = "2009",
month = "11",
doi = "10.1109/TVCG.2009.164",
language = "English (US)",
volume = "15",
pages = "1563--1570",
journal = "IEEE Transactions on Visualization and Computer Graphics",
issn = "1077-2626",
publisher = "IEEE Computer Society",
number = "6",

}

TY - JOUR

T1 - Mapping high-fidelity volume rendering for medical imaging to CPU, GPU and many-core architectures

AU - Smelyanskiy, Mikhail

AU - Holmes III, David R.

AU - Chhugani, Jatin

AU - Larson, Alan

AU - Carmean, Douglas M.

AU - Hanson, Dennis

AU - Dubey, Pradeep

AU - Augustine, Kurt

AU - Kim, Daehyun

AU - Kyker, Alan

AU - Lee, Victor W.

AU - Nguyen, Anthony D.

AU - Seiler, Larry

AU - Robb, Richard

PY - 2009/11

Y1 - 2009/11

N2 - Medical volumetric imaging requires high fidelity, high performance rendering algorithms. We motivate and analyze new volumetric rendering algorithms that are suited to modern parallel processing architectures. First, we describe the three major categories of volume rendering algorithms and confirm through an imaging scientist-guided evaluation that ray-casting is the most acceptable. We describe a thread- and data-parallel implementation of ray-casting that makes it amenable to key architectural trends of three modern commodity parallel architectures: multi-core, GPU, and an upcoming many-y-core Intel R architecture code-named Larrabee. We achieve more than an order of magnitude performance improvement on a number of large 3D medical datasets. We further describe a data compression scheme that significantly reduces data-transfer overhead. This allows our approach to scale well to large numbers of Larrabee cores.

AB - Medical volumetric imaging requires high fidelity, high performance rendering algorithms. We motivate and analyze new volumetric rendering algorithms that are suited to modern parallel processing architectures. First, we describe the three major categories of volume rendering algorithms and confirm through an imaging scientist-guided evaluation that ray-casting is the most acceptable. We describe a thread- and data-parallel implementation of ray-casting that makes it amenable to key architectural trends of three modern commodity parallel architectures: multi-core, GPU, and an upcoming many-y-core Intel R architecture code-named Larrabee. We achieve more than an order of magnitude performance improvement on a number of large 3D medical datasets. We further describe a data compression scheme that significantly reduces data-transfer overhead. This allows our approach to scale well to large numbers of Larrabee cores.

KW - GPGPU

KW - Graphics Architecture

KW - Many-core Computing

KW - Medical Imaging

KW - Parallel Processing

KW - Volume Compositing

UR - http://www.scopus.com/inward/record.url?scp=70350681243&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350681243&partnerID=8YFLogxK

U2 - 10.1109/TVCG.2009.164

DO - 10.1109/TVCG.2009.164

M3 - Article

C2 - 19834234

AN - SCOPUS:70350681243

VL - 15

SP - 1563

EP - 1570

JO - IEEE Transactions on Visualization and Computer Graphics

JF - IEEE Transactions on Visualization and Computer Graphics

SN - 1077-2626

IS - 6

M1 - 5290774

ER -