iSTAPLE: Improved label fusion for segmentation by combining STAPLE with image intensity

Xiaofeng Liu, Albert Montillo, Ek T. Tan, John F. Schenck

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

8 Citations (Scopus)

Abstract

Multi-atlas based methods have been a trend for robust and automated image segmentation. In general these methods first transfer prior manual segmentations, i.e., label maps, on a set of atlases to a given target image through image registration. These multiple label maps are then fused together to produce segmentations of the target image through voting strategy or statistical fusing, e.g., STAPLE. STAPLE simultaneously estimates the true segmentation and the label map performance level, but has been shown inaccurate for multi-atlas segmentation because it is determined completely on the propagated label maps without considering the target image intensity. We develop a new method, called iSTAPLE, that combines target image intensity into a similar maximum likelihood estimate (MLE) framework as in STAPLE to take advantage of both intensity-based segmentation and statistical label fusion based on atlas consensus and performance level. The MLE framework is then solved using a modified EM algorithm to simultaneously estimate the intensity profiles of structures of interest as well as the true segmentation and atlas performance level. Unlike other methods, iSTAPLE does not require the target image to have same image contrast and intensity range as the atlas images, which greatly extends the use of atlases. Experiments on whole brain segmentation showed that iSTAPLE performed consistently better than STAPLE.

Original languageEnglish (US)
Title of host publicationMedical Imaging 2013
Subtitle of host publicationImage Processing
Volume8669
DOIs
StatePublished - Jun 3 2013
EventMedical Imaging 2013: Image Processing - Lake Buena Vista, FL, United States
Duration: Feb 10 2013Feb 12 2013

Other

OtherMedical Imaging 2013: Image Processing
CountryUnited States
CityLake Buena Vista, FL
Period2/10/132/12/13

Fingerprint

Atlases
Labels
Fusion reactions
fusion
Likelihood Functions
Maximum likelihood
maximum likelihood estimates
Image registration
Image segmentation
Brain
voting
Politics
image contrast
estimates
brain
trends
Experiments
profiles

Keywords

  • Brain segmentation
  • Label fusion
  • STAPLE

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Atomic and Molecular Physics, and Optics
  • Biomaterials
  • Radiology Nuclear Medicine and imaging

Cite this

Liu, X., Montillo, A., Tan, E. T., & Schenck, J. F. (2013). iSTAPLE: Improved label fusion for segmentation by combining STAPLE with image intensity. In Medical Imaging 2013: Image Processing (Vol. 8669). [86692O] https://doi.org/10.1117/12.2006447

iSTAPLE : Improved label fusion for segmentation by combining STAPLE with image intensity. / Liu, Xiaofeng; Montillo, Albert; Tan, Ek T.; Schenck, John F.

Medical Imaging 2013: Image Processing. Vol. 8669 2013. 86692O.

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

Liu, X, Montillo, A, Tan, ET & Schenck, JF 2013, iSTAPLE: Improved label fusion for segmentation by combining STAPLE with image intensity. in Medical Imaging 2013: Image Processing. vol. 8669, 86692O, Medical Imaging 2013: Image Processing, Lake Buena Vista, FL, United States, 2/10/13. https://doi.org/10.1117/12.2006447
Liu X, Montillo A, Tan ET, Schenck JF. iSTAPLE: Improved label fusion for segmentation by combining STAPLE with image intensity. In Medical Imaging 2013: Image Processing. Vol. 8669. 2013. 86692O https://doi.org/10.1117/12.2006447
Liu, Xiaofeng ; Montillo, Albert ; Tan, Ek T. ; Schenck, John F. / iSTAPLE : Improved label fusion for segmentation by combining STAPLE with image intensity. Medical Imaging 2013: Image Processing. Vol. 8669 2013.
@inproceedings{d44d00597aad4ae3a80b6d8c5f3cbd5e,
title = "iSTAPLE: Improved label fusion for segmentation by combining STAPLE with image intensity",
abstract = "Multi-atlas based methods have been a trend for robust and automated image segmentation. In general these methods first transfer prior manual segmentations, i.e., label maps, on a set of atlases to a given target image through image registration. These multiple label maps are then fused together to produce segmentations of the target image through voting strategy or statistical fusing, e.g., STAPLE. STAPLE simultaneously estimates the true segmentation and the label map performance level, but has been shown inaccurate for multi-atlas segmentation because it is determined completely on the propagated label maps without considering the target image intensity. We develop a new method, called iSTAPLE, that combines target image intensity into a similar maximum likelihood estimate (MLE) framework as in STAPLE to take advantage of both intensity-based segmentation and statistical label fusion based on atlas consensus and performance level. The MLE framework is then solved using a modified EM algorithm to simultaneously estimate the intensity profiles of structures of interest as well as the true segmentation and atlas performance level. Unlike other methods, iSTAPLE does not require the target image to have same image contrast and intensity range as the atlas images, which greatly extends the use of atlases. Experiments on whole brain segmentation showed that iSTAPLE performed consistently better than STAPLE.",
keywords = "Brain segmentation, Label fusion, STAPLE",
author = "Xiaofeng Liu and Albert Montillo and Tan, {Ek T.} and Schenck, {John F.}",
year = "2013",
month = "6",
day = "3",
doi = "10.1117/12.2006447",
language = "English (US)",
isbn = "9780819494436",
volume = "8669",
booktitle = "Medical Imaging 2013",

}

TY - GEN

T1 - iSTAPLE

T2 - Improved label fusion for segmentation by combining STAPLE with image intensity

AU - Liu, Xiaofeng

AU - Montillo, Albert

AU - Tan, Ek T.

AU - Schenck, John F.

PY - 2013/6/3

Y1 - 2013/6/3

N2 - Multi-atlas based methods have been a trend for robust and automated image segmentation. In general these methods first transfer prior manual segmentations, i.e., label maps, on a set of atlases to a given target image through image registration. These multiple label maps are then fused together to produce segmentations of the target image through voting strategy or statistical fusing, e.g., STAPLE. STAPLE simultaneously estimates the true segmentation and the label map performance level, but has been shown inaccurate for multi-atlas segmentation because it is determined completely on the propagated label maps without considering the target image intensity. We develop a new method, called iSTAPLE, that combines target image intensity into a similar maximum likelihood estimate (MLE) framework as in STAPLE to take advantage of both intensity-based segmentation and statistical label fusion based on atlas consensus and performance level. The MLE framework is then solved using a modified EM algorithm to simultaneously estimate the intensity profiles of structures of interest as well as the true segmentation and atlas performance level. Unlike other methods, iSTAPLE does not require the target image to have same image contrast and intensity range as the atlas images, which greatly extends the use of atlases. Experiments on whole brain segmentation showed that iSTAPLE performed consistently better than STAPLE.

AB - Multi-atlas based methods have been a trend for robust and automated image segmentation. In general these methods first transfer prior manual segmentations, i.e., label maps, on a set of atlases to a given target image through image registration. These multiple label maps are then fused together to produce segmentations of the target image through voting strategy or statistical fusing, e.g., STAPLE. STAPLE simultaneously estimates the true segmentation and the label map performance level, but has been shown inaccurate for multi-atlas segmentation because it is determined completely on the propagated label maps without considering the target image intensity. We develop a new method, called iSTAPLE, that combines target image intensity into a similar maximum likelihood estimate (MLE) framework as in STAPLE to take advantage of both intensity-based segmentation and statistical label fusion based on atlas consensus and performance level. The MLE framework is then solved using a modified EM algorithm to simultaneously estimate the intensity profiles of structures of interest as well as the true segmentation and atlas performance level. Unlike other methods, iSTAPLE does not require the target image to have same image contrast and intensity range as the atlas images, which greatly extends the use of atlases. Experiments on whole brain segmentation showed that iSTAPLE performed consistently better than STAPLE.

KW - Brain segmentation

KW - Label fusion

KW - STAPLE

UR - http://www.scopus.com/inward/record.url?scp=84878305216&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84878305216&partnerID=8YFLogxK

U2 - 10.1117/12.2006447

DO - 10.1117/12.2006447

M3 - Conference contribution

C2 - 31741552

AN - SCOPUS:84878305216

SN - 9780819494436

VL - 8669

BT - Medical Imaging 2013

ER -