FastGeodis: Fast Generalised Geodesic Distance Transform

Muhammad Asad; Reuben Dorent; Tom Vercauteren

doi:10.21105/joss.04532

FastGeodis: Fast Generalised Geodesic Distance Transform

Muhammad Asad, Reuben Dorent, Tom Vercauteren

Surgical & Interventional Engineering

Research output: Contribution to journal › Article › peer-review

53 Downloads (Pure)

Abstract

The FastGeodis package provides an efficient implementation for computing Geodesic and Euclidean distance transforms (or a mixture of both), targeting efficient utilisation of CPU and GPU hardware. In particular, it implements the paralellisable raster scan method from Criminisi et al. (2009), where elements in a row (2D) or plane (3D) can be computed with parallel threads. This package is able to handle 2D as well as 3D data, where it achieves up to a 20x speedup on a CPU and up to a 74x speedup on a GPU as compared to an existing open-source library (Wang, 2020) that uses a non-parallelisable single-thread CPU implementation. The performance speedups reported here were evaluated using 3D volume data on an Nvidia GeForce Titan X (12 GB) with a 6-Core Intel Xeon E5-1650 CPU. Further in-depth comparison of performance improvements are discussed in the FastGeodis documentation: https://fastgeodis.readthedocs.io

Original language	Undefined/Unknown
Journal	Journal of Open Source Software
DOIs	https://doi.org/10.21105/joss.04532
Publication status	Published - 26 Jul 2022

Keywords

cs.CV
eess.IV

Access to Document

10.21105/joss.04532

2208.00001v2

Cite this

@article{41f41029da9b4b6391551a4935d6d811,

title = "FastGeodis: Fast Generalised Geodesic Distance Transform",

abstract = "The FastGeodis package provides an efficient implementation for computing Geodesic and Euclidean distance transforms (or a mixture of both), targeting efficient utilisation of CPU and GPU hardware. In particular, it implements the paralellisable raster scan method from Criminisi et al. (2009), where elements in a row (2D) or plane (3D) can be computed with parallel threads. This package is able to handle 2D as well as 3D data, where it achieves up to a 20x speedup on a CPU and up to a 74x speedup on a GPU as compared to an existing open-source library (Wang, 2020) that uses a non-parallelisable single-thread CPU implementation. The performance speedups reported here were evaluated using 3D volume data on an Nvidia GeForce Titan X (12 GB) with a 6-Core Intel Xeon E5-1650 CPU. Further in-depth comparison of performance improvements are discussed in the FastGeodis documentation: https://fastgeodis.readthedocs.io",

keywords = "cs.CV, eess.IV",

author = "Muhammad Asad and Reuben Dorent and Tom Vercauteren",

note = "Accepted at Journal of Open Source Software (JOSS)",

year = "2022",

month = jul,

day = "26",

doi = "10.21105/joss.04532",

language = "Undefined/Unknown",

journal = "Journal of Open Source Software",

issn = "2475-9066",

publisher = "Open Journals",

}

TY - JOUR

T1 - FastGeodis

T2 - Fast Generalised Geodesic Distance Transform

AU - Asad, Muhammad

AU - Dorent, Reuben

AU - Vercauteren, Tom

N1 - Accepted at Journal of Open Source Software (JOSS)

PY - 2022/7/26

Y1 - 2022/7/26

N2 - The FastGeodis package provides an efficient implementation for computing Geodesic and Euclidean distance transforms (or a mixture of both), targeting efficient utilisation of CPU and GPU hardware. In particular, it implements the paralellisable raster scan method from Criminisi et al. (2009), where elements in a row (2D) or plane (3D) can be computed with parallel threads. This package is able to handle 2D as well as 3D data, where it achieves up to a 20x speedup on a CPU and up to a 74x speedup on a GPU as compared to an existing open-source library (Wang, 2020) that uses a non-parallelisable single-thread CPU implementation. The performance speedups reported here were evaluated using 3D volume data on an Nvidia GeForce Titan X (12 GB) with a 6-Core Intel Xeon E5-1650 CPU. Further in-depth comparison of performance improvements are discussed in the FastGeodis documentation: https://fastgeodis.readthedocs.io

AB - The FastGeodis package provides an efficient implementation for computing Geodesic and Euclidean distance transforms (or a mixture of both), targeting efficient utilisation of CPU and GPU hardware. In particular, it implements the paralellisable raster scan method from Criminisi et al. (2009), where elements in a row (2D) or plane (3D) can be computed with parallel threads. This package is able to handle 2D as well as 3D data, where it achieves up to a 20x speedup on a CPU and up to a 74x speedup on a GPU as compared to an existing open-source library (Wang, 2020) that uses a non-parallelisable single-thread CPU implementation. The performance speedups reported here were evaluated using 3D volume data on an Nvidia GeForce Titan X (12 GB) with a 6-Core Intel Xeon E5-1650 CPU. Further in-depth comparison of performance improvements are discussed in the FastGeodis documentation: https://fastgeodis.readthedocs.io

KW - cs.CV

KW - eess.IV

U2 - 10.21105/joss.04532

DO - 10.21105/joss.04532

M3 - Article

SN - 2475-9066

JO - Journal of Open Source Software

JF - Journal of Open Source Software

ER -