Incorporating priors for medical image segmentation using a genetic algorithm

Payel Ghosh, Melanie Mitchell, James Tanyi, Arthur Hung

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

Medical image segmentation is typically performed manually by a physician to delineate gross tumor volumes for treatment planning and diagnosis. Manual segmentation is performed by medical experts using prior knowledge of organ shapes and locations but is prone to reader subjectivity and inconsistency. Automating the process is challenging due to poor tissue contrast and ill-defined organ/tissue boundaries in medical images. This paper presents a genetic algorithm for combining representations of learned information such as known shapes, regional properties and relative position of objects into a single framework to perform automated three-dimensional segmentation. The algorithm has been tested for prostate segmentation on pelvic computed tomography and magnetic resonance images.

Original languageEnglish (US)
Pages (from-to)181-194
Number of pages14
JournalNeurocomputing
Volume195
DOIs
StatePublished - Jun 26 2016

Fingerprint

Image segmentation
Genetic algorithms
Tissue
Magnetic resonance
Tumor Burden
Tomography
Tumors
Prostate
Magnetic Resonance Spectroscopy
Physicians
Planning
Therapeutics

Keywords

  • Genetic Algorithms
  • Medical image segmentation
  • Relative position priors
  • Shape Modeling
  • Texture Analysis

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Cognitive Neuroscience

Cite this

Incorporating priors for medical image segmentation using a genetic algorithm. / Ghosh, Payel; Mitchell, Melanie; Tanyi, James; Hung, Arthur.

In: Neurocomputing, Vol. 195, 26.06.2016, p. 181-194.

Research output: Contribution to journalArticle

@article{a481b1b0986f46f08b9f3fec640bb2f4,
title = "Incorporating priors for medical image segmentation using a genetic algorithm",
abstract = "Medical image segmentation is typically performed manually by a physician to delineate gross tumor volumes for treatment planning and diagnosis. Manual segmentation is performed by medical experts using prior knowledge of organ shapes and locations but is prone to reader subjectivity and inconsistency. Automating the process is challenging due to poor tissue contrast and ill-defined organ/tissue boundaries in medical images. This paper presents a genetic algorithm for combining representations of learned information such as known shapes, regional properties and relative position of objects into a single framework to perform automated three-dimensional segmentation. The algorithm has been tested for prostate segmentation on pelvic computed tomography and magnetic resonance images.",
keywords = "Genetic Algorithms, Medical image segmentation, Relative position priors, Shape Modeling, Texture Analysis",
author = "Payel Ghosh and Melanie Mitchell and James Tanyi and Arthur Hung",
year = "2016",
month = "6",
day = "26",
doi = "10.1016/j.neucom.2015.09.123",
language = "English (US)",
volume = "195",
pages = "181--194",
journal = "Neurocomputing",
issn = "0925-2312",
publisher = "Elsevier",

}

TY - JOUR

T1 - Incorporating priors for medical image segmentation using a genetic algorithm

AU - Ghosh, Payel

AU - Mitchell, Melanie

AU - Tanyi, James

AU - Hung, Arthur

PY - 2016/6/26

Y1 - 2016/6/26

N2 - Medical image segmentation is typically performed manually by a physician to delineate gross tumor volumes for treatment planning and diagnosis. Manual segmentation is performed by medical experts using prior knowledge of organ shapes and locations but is prone to reader subjectivity and inconsistency. Automating the process is challenging due to poor tissue contrast and ill-defined organ/tissue boundaries in medical images. This paper presents a genetic algorithm for combining representations of learned information such as known shapes, regional properties and relative position of objects into a single framework to perform automated three-dimensional segmentation. The algorithm has been tested for prostate segmentation on pelvic computed tomography and magnetic resonance images.

AB - Medical image segmentation is typically performed manually by a physician to delineate gross tumor volumes for treatment planning and diagnosis. Manual segmentation is performed by medical experts using prior knowledge of organ shapes and locations but is prone to reader subjectivity and inconsistency. Automating the process is challenging due to poor tissue contrast and ill-defined organ/tissue boundaries in medical images. This paper presents a genetic algorithm for combining representations of learned information such as known shapes, regional properties and relative position of objects into a single framework to perform automated three-dimensional segmentation. The algorithm has been tested for prostate segmentation on pelvic computed tomography and magnetic resonance images.

KW - Genetic Algorithms

KW - Medical image segmentation

KW - Relative position priors

KW - Shape Modeling

KW - Texture Analysis

UR - http://www.scopus.com/inward/record.url?scp=84964001423&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84964001423&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2015.09.123

DO - 10.1016/j.neucom.2015.09.123

M3 - Article

VL - 195

SP - 181

EP - 194

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

ER -