Compressed Sensing MRI Reconstruction on Intel HARPv2

Yushan Su, Michael Anderson, Jonathan I. Tamir, Michael Lustig, Kai Li

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Implementing the Iterative Soft-Thresholding Algorithm (ISTA) of compressed sensing for MRI image reconstruction is a good candidate for designing accelerators because real-Time functional MRI applications require intensive computations. A straightforward mapping of the computation graph of ISTA onto an FPGA, with a wide enough datapath to saturate memory bandwidth, would require substantial resources, such that a modest size FPGA would not fit the reconstruction pipeline for an entire MRI image. This paper proposes several methods to design the kernel components of ISTA, such as matrix transpose, datapath reuse, parallelism within maps, and data buffering to overcome the problem. Our implementation with Intel OpenCL SDK and performance evaluation on Intel HARPv2 show that our methods can map the reconstruction for the entire 256x256 MRI image with 8 or more channels to its FPGA, while achieving good overall performance.

Original languageEnglish (US)
Title of host publicationProceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages254-257
Number of pages4
ISBN (Electronic)9781728111315
DOIs
StatePublished - Apr 1 2019
Event27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019 - San Diego, United States
Duration: Apr 28 2019May 1 2019

Publication series

NameProceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

Conference

Conference27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019
CountryUnited States
CitySan Diego
Period4/28/195/1/19

Fingerprint

Compressed sensing
Magnetic resonance imaging
Field programmable gate arrays (FPGA)
Image reconstruction
Particle accelerators
Pipelines
Bandwidth
Data storage equipment

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Hardware and Architecture

Cite this

Su, Y., Anderson, M., Tamir, J. I., Lustig, M., & Li, K. (2019). Compressed Sensing MRI Reconstruction on Intel HARPv2. In Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019 (pp. 254-257). [8735528] (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/FCCM.2019.00041
Su, Yushan ; Anderson, Michael ; Tamir, Jonathan I. ; Lustig, Michael ; Li, Kai. / Compressed Sensing MRI Reconstruction on Intel HARPv2. Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 254-257 (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019).
@inproceedings{41fad81a2a544604941a5fa70a25f024,
title = "Compressed Sensing MRI Reconstruction on Intel HARPv2",
abstract = "Implementing the Iterative Soft-Thresholding Algorithm (ISTA) of compressed sensing for MRI image reconstruction is a good candidate for designing accelerators because real-Time functional MRI applications require intensive computations. A straightforward mapping of the computation graph of ISTA onto an FPGA, with a wide enough datapath to saturate memory bandwidth, would require substantial resources, such that a modest size FPGA would not fit the reconstruction pipeline for an entire MRI image. This paper proposes several methods to design the kernel components of ISTA, such as matrix transpose, datapath reuse, parallelism within maps, and data buffering to overcome the problem. Our implementation with Intel OpenCL SDK and performance evaluation on Intel HARPv2 show that our methods can map the reconstruction for the entire 256x256 MRI image with 8 or more channels to its FPGA, while achieving good overall performance.",
author = "Yushan Su and Michael Anderson and Tamir, {Jonathan I.} and Michael Lustig and Kai Li",
year = "2019",
month = "4",
day = "1",
doi = "10.1109/FCCM.2019.00041",
language = "English (US)",
series = "Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "254--257",
booktitle = "Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019",
address = "United States",

}

Su, Y, Anderson, M, Tamir, JI, Lustig, M & Li, K 2019, Compressed Sensing MRI Reconstruction on Intel HARPv2. in Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019., 8735528, Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019, Institute of Electrical and Electronics Engineers Inc., pp. 254-257, 27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019, San Diego, United States, 4/28/19. https://doi.org/10.1109/FCCM.2019.00041

Compressed Sensing MRI Reconstruction on Intel HARPv2. / Su, Yushan; Anderson, Michael; Tamir, Jonathan I.; Lustig, Michael; Li, Kai.

Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 254-257 8735528 (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Compressed Sensing MRI Reconstruction on Intel HARPv2

AU - Su, Yushan

AU - Anderson, Michael

AU - Tamir, Jonathan I.

AU - Lustig, Michael

AU - Li, Kai

PY - 2019/4/1

Y1 - 2019/4/1

N2 - Implementing the Iterative Soft-Thresholding Algorithm (ISTA) of compressed sensing for MRI image reconstruction is a good candidate for designing accelerators because real-Time functional MRI applications require intensive computations. A straightforward mapping of the computation graph of ISTA onto an FPGA, with a wide enough datapath to saturate memory bandwidth, would require substantial resources, such that a modest size FPGA would not fit the reconstruction pipeline for an entire MRI image. This paper proposes several methods to design the kernel components of ISTA, such as matrix transpose, datapath reuse, parallelism within maps, and data buffering to overcome the problem. Our implementation with Intel OpenCL SDK and performance evaluation on Intel HARPv2 show that our methods can map the reconstruction for the entire 256x256 MRI image with 8 or more channels to its FPGA, while achieving good overall performance.

AB - Implementing the Iterative Soft-Thresholding Algorithm (ISTA) of compressed sensing for MRI image reconstruction is a good candidate for designing accelerators because real-Time functional MRI applications require intensive computations. A straightforward mapping of the computation graph of ISTA onto an FPGA, with a wide enough datapath to saturate memory bandwidth, would require substantial resources, such that a modest size FPGA would not fit the reconstruction pipeline for an entire MRI image. This paper proposes several methods to design the kernel components of ISTA, such as matrix transpose, datapath reuse, parallelism within maps, and data buffering to overcome the problem. Our implementation with Intel OpenCL SDK and performance evaluation on Intel HARPv2 show that our methods can map the reconstruction for the entire 256x256 MRI image with 8 or more channels to its FPGA, while achieving good overall performance.

UR - http://www.scopus.com/inward/record.url?scp=85068350660&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068350660&partnerID=8YFLogxK

U2 - 10.1109/FCCM.2019.00041

DO - 10.1109/FCCM.2019.00041

M3 - Conference contribution

T3 - Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

SP - 254

EP - 257

BT - Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Su Y, Anderson M, Tamir JI, Lustig M, Li K. Compressed Sensing MRI Reconstruction on Intel HARPv2. In Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 254-257. 8735528. (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019). https://doi.org/10.1109/FCCM.2019.00041