Multi-object geodesic active contours (MOGAC).

Blake C. Lucas, Michael Kazhdan, Russell H Taylor

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

An emerging topic is to build image segmentation systems that can segment hundreds to thousands of objects (i.e. cell segmentation\tracking, full brain parcellation, full body segmentation, etc.). Multi-object Level Set Methods (MLSM) perform this task with the benefit of sub-pixel precision. However, current implementations of MLSM are not as computationally or memory efficient as their region growing and graph cut counterparts which lack sub-pixel precision. To address this performance gap, we present a novel parallel implementation of MLSM that leverages the sparse properties of the algorithm to minimize its memory footprint for multiple objects. The new method, Multi-Object Geodesic Active Contours (MOGAC), can represent N objects with just two functions: a label mask image and unsigned distance field. The time complexity of the algorithm is shown to be O((M (power)d)/P) for M (power)d pixels and P processing units in dimension d = {2,3}, independent of the number of objects. Results are presented for 2D and 3D image segmentation problems.

Original languageEnglish (US)
Title of host publicationMedical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
Pages404-412
Number of pages9
Volume15
EditionPt 2
StatePublished - 2012

Fingerprint

Cell Tracking
Masks
Brain
Power (Psychology)

ASJC Scopus subject areas

  • Medicine(all)

Cite this

Lucas, B. C., Kazhdan, M., & Taylor, R. H. (2012). Multi-object geodesic active contours (MOGAC). In Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention (Pt 2 ed., Vol. 15, pp. 404-412)

Multi-object geodesic active contours (MOGAC). / Lucas, Blake C.; Kazhdan, Michael; Taylor, Russell H.

Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. Vol. 15 Pt 2. ed. 2012. p. 404-412.

Research output: Chapter in Book/Report/Conference proceedingChapter

Lucas, BC, Kazhdan, M & Taylor, RH 2012, Multi-object geodesic active contours (MOGAC). in Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. Pt 2 edn, vol. 15, pp. 404-412.
Lucas BC, Kazhdan M, Taylor RH. Multi-object geodesic active contours (MOGAC). In Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. Pt 2 ed. Vol. 15. 2012. p. 404-412
Lucas, Blake C. ; Kazhdan, Michael ; Taylor, Russell H. / Multi-object geodesic active contours (MOGAC). Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. Vol. 15 Pt 2. ed. 2012. pp. 404-412
@inbook{d98b3d6a65ac425bb5bd9f087300ebbb,
title = "Multi-object geodesic active contours (MOGAC).",
abstract = "An emerging topic is to build image segmentation systems that can segment hundreds to thousands of objects (i.e. cell segmentation\tracking, full brain parcellation, full body segmentation, etc.). Multi-object Level Set Methods (MLSM) perform this task with the benefit of sub-pixel precision. However, current implementations of MLSM are not as computationally or memory efficient as their region growing and graph cut counterparts which lack sub-pixel precision. To address this performance gap, we present a novel parallel implementation of MLSM that leverages the sparse properties of the algorithm to minimize its memory footprint for multiple objects. The new method, Multi-Object Geodesic Active Contours (MOGAC), can represent N objects with just two functions: a label mask image and unsigned distance field. The time complexity of the algorithm is shown to be O((M (power)d)/P) for M (power)d pixels and P processing units in dimension d = {2,3}, independent of the number of objects. Results are presented for 2D and 3D image segmentation problems.",
author = "Lucas, {Blake C.} and Michael Kazhdan and Taylor, {Russell H}",
year = "2012",
language = "English (US)",
volume = "15",
pages = "404--412",
booktitle = "Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention",
edition = "Pt 2",

}

TY - CHAP

T1 - Multi-object geodesic active contours (MOGAC).

AU - Lucas, Blake C.

AU - Kazhdan, Michael

AU - Taylor, Russell H

PY - 2012

Y1 - 2012

N2 - An emerging topic is to build image segmentation systems that can segment hundreds to thousands of objects (i.e. cell segmentation\tracking, full brain parcellation, full body segmentation, etc.). Multi-object Level Set Methods (MLSM) perform this task with the benefit of sub-pixel precision. However, current implementations of MLSM are not as computationally or memory efficient as their region growing and graph cut counterparts which lack sub-pixel precision. To address this performance gap, we present a novel parallel implementation of MLSM that leverages the sparse properties of the algorithm to minimize its memory footprint for multiple objects. The new method, Multi-Object Geodesic Active Contours (MOGAC), can represent N objects with just two functions: a label mask image and unsigned distance field. The time complexity of the algorithm is shown to be O((M (power)d)/P) for M (power)d pixels and P processing units in dimension d = {2,3}, independent of the number of objects. Results are presented for 2D and 3D image segmentation problems.

AB - An emerging topic is to build image segmentation systems that can segment hundreds to thousands of objects (i.e. cell segmentation\tracking, full brain parcellation, full body segmentation, etc.). Multi-object Level Set Methods (MLSM) perform this task with the benefit of sub-pixel precision. However, current implementations of MLSM are not as computationally or memory efficient as their region growing and graph cut counterparts which lack sub-pixel precision. To address this performance gap, we present a novel parallel implementation of MLSM that leverages the sparse properties of the algorithm to minimize its memory footprint for multiple objects. The new method, Multi-Object Geodesic Active Contours (MOGAC), can represent N objects with just two functions: a label mask image and unsigned distance field. The time complexity of the algorithm is shown to be O((M (power)d)/P) for M (power)d pixels and P processing units in dimension d = {2,3}, independent of the number of objects. Results are presented for 2D and 3D image segmentation problems.

UR - http://www.scopus.com/inward/record.url?scp=84872900369&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84872900369&partnerID=8YFLogxK

M3 - Chapter

C2 - 23286074

AN - SCOPUS:84872900369

VL - 15

SP - 404

EP - 412

BT - Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention

ER -