Integrating an adaptive region-based appearance model with a landmark-free statistical shape model: Application to prostate MRI segmentation

Robert Toth, Julie Bulman, Amish D. Patel, B. Nicholas Bloch, Elizabeth M. Genega, Neil M. Rofsky, Robert E. Lenkinski, Anant Madabhushi

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

7 Citations (Scopus)

Abstract

In this paper we present a system for segmenting medical images using statistical shape models (SSM's) which is landmark free, fully 3D, and accurate. To overcome the limitations associated with previous 3D landmark-based SSM's, our system creates a levelset-based SSM which uses the minimum distance from each voxel in the image to the object's surface to define a shape. Subsequently, an advanced statistical appearance model (SAM) is generated to model the object of interest. This SAM is based on a series of statistical texture features calculated from each image, modeled by a Gaussian Mixture Model. In order to segment the object of interest on a new image, a Bayesian classifier is first employed to pre-classify the image voxels as belonging to the foreground object of interest or the background. The result of the Bayesian classifier is then employed for optimally fitting the SSM so there is maximum agreement between the SAM and the SSM. The SAM is then able to adaptively learn the statistics of the textures of the foreground and background voxels on the new image. The fitting of the SSM, and the adaptive updating of the SAM is repeated until convergence. We have tested our system on 36 T2-w, 3.0 Tesla, in vivo, endorectal prostate images. The results showed that our system achieves a Dice similarity coefficient of.84±.04, with a median Dice value of.86, which is comparable (and in most cases superior) to other state of the art prostate segmentation systems. Further, unlike most other state of the art prostate segmentation schemes, our scheme is fully automated requiring no user intervention.

Original languageEnglish (US)
Title of host publicationProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume7962
DOIs
StatePublished - 2011
EventMedical Imaging 2011: Image Processing - Lake Buena Vista, FL, United States
Duration: Feb 14 2011Feb 16 2011

Other

OtherMedical Imaging 2011: Image Processing
CountryUnited States
CityLake Buena Vista, FL
Period2/14/112/16/11

Fingerprint

landmarks
Statistical Models
Magnetic resonance imaging
Prostate
Classifiers
Textures
classifiers
textures
Statistics

Keywords

  • Active Appearance Model
  • Active Shape Model
  • Levelset
  • Prostate MRI
  • Prostate Segmentation

ASJC Scopus subject areas

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

Cite this

Toth, R., Bulman, J., Patel, A. D., Bloch, B. N., Genega, E. M., Rofsky, N. M., ... Madabhushi, A. (2011). Integrating an adaptive region-based appearance model with a landmark-free statistical shape model: Application to prostate MRI segmentation. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE (Vol. 7962). [79622V] https://doi.org/10.1117/12.878346

Integrating an adaptive region-based appearance model with a landmark-free statistical shape model : Application to prostate MRI segmentation. / Toth, Robert; Bulman, Julie; Patel, Amish D.; Bloch, B. Nicholas; Genega, Elizabeth M.; Rofsky, Neil M.; Lenkinski, Robert E.; Madabhushi, Anant.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 7962 2011. 79622V.

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

Toth, R, Bulman, J, Patel, AD, Bloch, BN, Genega, EM, Rofsky, NM, Lenkinski, RE & Madabhushi, A 2011, Integrating an adaptive region-based appearance model with a landmark-free statistical shape model: Application to prostate MRI segmentation. in Progress in Biomedical Optics and Imaging - Proceedings of SPIE. vol. 7962, 79622V, Medical Imaging 2011: Image Processing, Lake Buena Vista, FL, United States, 2/14/11. https://doi.org/10.1117/12.878346
Toth R, Bulman J, Patel AD, Bloch BN, Genega EM, Rofsky NM et al. Integrating an adaptive region-based appearance model with a landmark-free statistical shape model: Application to prostate MRI segmentation. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 7962. 2011. 79622V https://doi.org/10.1117/12.878346
Toth, Robert ; Bulman, Julie ; Patel, Amish D. ; Bloch, B. Nicholas ; Genega, Elizabeth M. ; Rofsky, Neil M. ; Lenkinski, Robert E. ; Madabhushi, Anant. / Integrating an adaptive region-based appearance model with a landmark-free statistical shape model : Application to prostate MRI segmentation. Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 7962 2011.
@inproceedings{79dd92b7546245d09e1ee098b5e6d13d,
title = "Integrating an adaptive region-based appearance model with a landmark-free statistical shape model: Application to prostate MRI segmentation",
abstract = "In this paper we present a system for segmenting medical images using statistical shape models (SSM's) which is landmark free, fully 3D, and accurate. To overcome the limitations associated with previous 3D landmark-based SSM's, our system creates a levelset-based SSM which uses the minimum distance from each voxel in the image to the object's surface to define a shape. Subsequently, an advanced statistical appearance model (SAM) is generated to model the object of interest. This SAM is based on a series of statistical texture features calculated from each image, modeled by a Gaussian Mixture Model. In order to segment the object of interest on a new image, a Bayesian classifier is first employed to pre-classify the image voxels as belonging to the foreground object of interest or the background. The result of the Bayesian classifier is then employed for optimally fitting the SSM so there is maximum agreement between the SAM and the SSM. The SAM is then able to adaptively learn the statistics of the textures of the foreground and background voxels on the new image. The fitting of the SSM, and the adaptive updating of the SAM is repeated until convergence. We have tested our system on 36 T2-w, 3.0 Tesla, in vivo, endorectal prostate images. The results showed that our system achieves a Dice similarity coefficient of.84±.04, with a median Dice value of.86, which is comparable (and in most cases superior) to other state of the art prostate segmentation systems. Further, unlike most other state of the art prostate segmentation schemes, our scheme is fully automated requiring no user intervention.",
keywords = "Active Appearance Model, Active Shape Model, Levelset, Prostate MRI, Prostate Segmentation",
author = "Robert Toth and Julie Bulman and Patel, {Amish D.} and Bloch, {B. Nicholas} and Genega, {Elizabeth M.} and Rofsky, {Neil M.} and Lenkinski, {Robert E.} and Anant Madabhushi",
year = "2011",
doi = "10.1117/12.878346",
language = "English (US)",
isbn = "9780819485045",
volume = "7962",
booktitle = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

}

TY - GEN

T1 - Integrating an adaptive region-based appearance model with a landmark-free statistical shape model

T2 - Application to prostate MRI segmentation

AU - Toth, Robert

AU - Bulman, Julie

AU - Patel, Amish D.

AU - Bloch, B. Nicholas

AU - Genega, Elizabeth M.

AU - Rofsky, Neil M.

AU - Lenkinski, Robert E.

AU - Madabhushi, Anant

PY - 2011

Y1 - 2011

N2 - In this paper we present a system for segmenting medical images using statistical shape models (SSM's) which is landmark free, fully 3D, and accurate. To overcome the limitations associated with previous 3D landmark-based SSM's, our system creates a levelset-based SSM which uses the minimum distance from each voxel in the image to the object's surface to define a shape. Subsequently, an advanced statistical appearance model (SAM) is generated to model the object of interest. This SAM is based on a series of statistical texture features calculated from each image, modeled by a Gaussian Mixture Model. In order to segment the object of interest on a new image, a Bayesian classifier is first employed to pre-classify the image voxels as belonging to the foreground object of interest or the background. The result of the Bayesian classifier is then employed for optimally fitting the SSM so there is maximum agreement between the SAM and the SSM. The SAM is then able to adaptively learn the statistics of the textures of the foreground and background voxels on the new image. The fitting of the SSM, and the adaptive updating of the SAM is repeated until convergence. We have tested our system on 36 T2-w, 3.0 Tesla, in vivo, endorectal prostate images. The results showed that our system achieves a Dice similarity coefficient of.84±.04, with a median Dice value of.86, which is comparable (and in most cases superior) to other state of the art prostate segmentation systems. Further, unlike most other state of the art prostate segmentation schemes, our scheme is fully automated requiring no user intervention.

AB - In this paper we present a system for segmenting medical images using statistical shape models (SSM's) which is landmark free, fully 3D, and accurate. To overcome the limitations associated with previous 3D landmark-based SSM's, our system creates a levelset-based SSM which uses the minimum distance from each voxel in the image to the object's surface to define a shape. Subsequently, an advanced statistical appearance model (SAM) is generated to model the object of interest. This SAM is based on a series of statistical texture features calculated from each image, modeled by a Gaussian Mixture Model. In order to segment the object of interest on a new image, a Bayesian classifier is first employed to pre-classify the image voxels as belonging to the foreground object of interest or the background. The result of the Bayesian classifier is then employed for optimally fitting the SSM so there is maximum agreement between the SAM and the SSM. The SAM is then able to adaptively learn the statistics of the textures of the foreground and background voxels on the new image. The fitting of the SSM, and the adaptive updating of the SAM is repeated until convergence. We have tested our system on 36 T2-w, 3.0 Tesla, in vivo, endorectal prostate images. The results showed that our system achieves a Dice similarity coefficient of.84±.04, with a median Dice value of.86, which is comparable (and in most cases superior) to other state of the art prostate segmentation systems. Further, unlike most other state of the art prostate segmentation schemes, our scheme is fully automated requiring no user intervention.

KW - Active Appearance Model

KW - Active Shape Model

KW - Levelset

KW - Prostate MRI

KW - Prostate Segmentation

UR - http://www.scopus.com/inward/record.url?scp=79958004465&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79958004465&partnerID=8YFLogxK

U2 - 10.1117/12.878346

DO - 10.1117/12.878346

M3 - Conference contribution

AN - SCOPUS:79958004465

SN - 9780819485045

VL - 7962

BT - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

ER -