Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation

Dwarikanath Mahapatra; Peter J. Schüffler; Jeroen A.W. Tielbeek; Carl A.J. Puylaert; Jesica C. Makanyanga; Alex Menys; Rado Andriantsimiavona; Jaap Stoker; Stuart A. Taylor; Franciscus M. Vos; Joachim M. Buhmann

doi:10.1007/978-3-319-13692-9_13

Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation

Dwarikanath Mahapatra, Peter J. Schüffler, Jeroen A.W. Tielbeek, Carl A.J. Puylaert, Jesica C. Makanyanga, Alex Menys, Rado Andriantsimiavona, Jaap Stoker, Stuart A. Taylor, Franciscus M. Vos, Joachim M. Buhmann

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

We propose a graph cut (GC) based approach for combining annotations from multiple experts and segmenting Crohns disease (CD) tissues in magnetic resonance (MR) images. Random forest (RF) based semi supervised learning (SSL) predicts missing expert labels while a novel self consistency (SC) score quantifies the reliability of each expert label and also serves as the penalty cost in a second order Markov random field (MRF) cost function. The final consensus label is obtained by GC optimization. Experimental results on synthetic images and real CD patient data show our final segmentation to be more accurate than those obtained by competing methods. It also highlights the effectiveness of SC score in quantifying expert reliability and accuracy of SSL in predicting missing labels.

Original language	English
Title of host publication	Abdominal Imaging
Subtitle of host publication	Computational and Clinical Applications - 6th International Workshop, ABDI 2014 held in conjunction with MICCAI 2014
Editors	Hiroyuki Yoshida, Janne J. Näppi, Sanjay Saini
Publisher	Springer Verlag
Pages	139-147
Number of pages	9
ISBN (Electronic)	9783319136912
DOIs	https://doi.org/10.1007/978-3-319-13692-9_13
State	Published - 2014
Externally published	Yes
Event	6th International Workshop on Abdominal Imaging: Computational and Clinical Applications, ABDI 2014 held in conjunction with 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014 - Cambridge, United States Duration: 14 Sep 2014 → 14 Sep 2014

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	8676
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	6th International Workshop on Abdominal Imaging: Computational and Clinical Applications, ABDI 2014 held in conjunction with 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014
Country/Territory	United States
City	Cambridge
Period	14/09/14 → 14/09/14

Keywords

Crohn’s disease
Graph cut
Magnetic resonance
Markov random field
Random forest

Access to Document

10.1007/978-3-319-13692-9_13

Cite this

Mahapatra, D., Schüffler, P. J., Tielbeek, J. A. W., Puylaert, C. A. J., Makanyanga, J. C., Menys, A., Andriantsimiavona, R., Stoker, J., Taylor, S. A., Vos, F. M., & Buhmann, J. M. (2014). Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation. In H. Yoshida, J. J. Näppi, & S. Saini (Eds.), Abdominal Imaging: Computational and Clinical Applications - 6th International Workshop, ABDI 2014 held in conjunction with MICCAI 2014 (pp. 139-147). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8676). Springer Verlag. https://doi.org/10.1007/978-3-319-13692-9_13

Mahapatra, Dwarikanath ; Schüffler, Peter J. ; Tielbeek, Jeroen A.W. et al. / Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation. Abdominal Imaging: Computational and Clinical Applications - 6th International Workshop, ABDI 2014 held in conjunction with MICCAI 2014. editor / Hiroyuki Yoshida ; Janne J. Näppi ; Sanjay Saini. Springer Verlag, 2014. pp. 139-147 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{99885b7061d34f2a87cf408dea2909cb,

title = "Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn{\textquoteright}s disease segmentation",

abstract = "We propose a graph cut (GC) based approach for combining annotations from multiple experts and segmenting Crohns disease (CD) tissues in magnetic resonance (MR) images. Random forest (RF) based semi supervised learning (SSL) predicts missing expert labels while a novel self consistency (SC) score quantifies the reliability of each expert label and also serves as the penalty cost in a second order Markov random field (MRF) cost function. The final consensus label is obtained by GC optimization. Experimental results on synthetic images and real CD patient data show our final segmentation to be more accurate than those obtained by competing methods. It also highlights the effectiveness of SC score in quantifying expert reliability and accuracy of SSL in predicting missing labels.",

keywords = "Crohn{\textquoteright}s disease, Graph cut, Magnetic resonance, Markov random field, Random forest",

author = "Dwarikanath Mahapatra and Sch{\"u}ffler, {Peter J.} and Tielbeek, {Jeroen A.W.} and Puylaert, {Carl A.J.} and Makanyanga, {Jesica C.} and Alex Menys and Rado Andriantsimiavona and Jaap Stoker and Taylor, {Stuart A.} and Vos, {Franciscus M.} and Buhmann, {Joachim M.}",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2014.; 6th International Workshop on Abdominal Imaging: Computational and Clinical Applications, ABDI 2014 held in conjunction with 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014 ; Conference date: 14-09-2014 Through 14-09-2014",

year = "2014",

doi = "10.1007/978-3-319-13692-9_13",

language = "English",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "139--147",

editor = "Hiroyuki Yoshida and N{\"a}ppi, {Janne J.} and Sanjay Saini",

booktitle = "Abdominal Imaging",

}

Mahapatra, D, Schüffler, PJ, Tielbeek, JAW, Puylaert, CAJ, Makanyanga, JC, Menys, A, Andriantsimiavona, R, Stoker, J, Taylor, SA, Vos, FM & Buhmann, JM 2014, Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation. in H Yoshida, JJ Näppi & S Saini (eds), Abdominal Imaging: Computational and Clinical Applications - 6th International Workshop, ABDI 2014 held in conjunction with MICCAI 2014. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8676, Springer Verlag, pp. 139-147, 6th International Workshop on Abdominal Imaging: Computational and Clinical Applications, ABDI 2014 held in conjunction with 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014, Cambridge, United States, 14/09/14. https://doi.org/10.1007/978-3-319-13692-9_13

Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation. / Mahapatra, Dwarikanath; Schüffler, Peter J.; Tielbeek, Jeroen A.W. et al.
Abdominal Imaging: Computational and Clinical Applications - 6th International Workshop, ABDI 2014 held in conjunction with MICCAI 2014. ed. / Hiroyuki Yoshida; Janne J. Näppi; Sanjay Saini. Springer Verlag, 2014. p. 139-147 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8676).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation

AU - Mahapatra, Dwarikanath

AU - Schüffler, Peter J.

AU - Tielbeek, Jeroen A.W.

AU - Puylaert, Carl A.J.

AU - Makanyanga, Jesica C.

AU - Menys, Alex

AU - Andriantsimiavona, Rado

AU - Stoker, Jaap

AU - Taylor, Stuart A.

AU - Vos, Franciscus M.

AU - Buhmann, Joachim M.

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2014.

PY - 2014

Y1 - 2014

N2 - We propose a graph cut (GC) based approach for combining annotations from multiple experts and segmenting Crohns disease (CD) tissues in magnetic resonance (MR) images. Random forest (RF) based semi supervised learning (SSL) predicts missing expert labels while a novel self consistency (SC) score quantifies the reliability of each expert label and also serves as the penalty cost in a second order Markov random field (MRF) cost function. The final consensus label is obtained by GC optimization. Experimental results on synthetic images and real CD patient data show our final segmentation to be more accurate than those obtained by competing methods. It also highlights the effectiveness of SC score in quantifying expert reliability and accuracy of SSL in predicting missing labels.

AB - We propose a graph cut (GC) based approach for combining annotations from multiple experts and segmenting Crohns disease (CD) tissues in magnetic resonance (MR) images. Random forest (RF) based semi supervised learning (SSL) predicts missing expert labels while a novel self consistency (SC) score quantifies the reliability of each expert label and also serves as the penalty cost in a second order Markov random field (MRF) cost function. The final consensus label is obtained by GC optimization. Experimental results on synthetic images and real CD patient data show our final segmentation to be more accurate than those obtained by competing methods. It also highlights the effectiveness of SC score in quantifying expert reliability and accuracy of SSL in predicting missing labels.

KW - Crohn’s disease

KW - Graph cut

KW - Magnetic resonance

KW - Markov random field

KW - Random forest

UR - http://www.scopus.com/inward/record.url?scp=84927669781&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-13692-9_13

DO - 10.1007/978-3-319-13692-9_13

M3 - Conference contribution

AN - SCOPUS:84927669781

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 139

EP - 147

BT - Abdominal Imaging

A2 - Yoshida, Hiroyuki

A2 - Näppi, Janne J.

A2 - Saini, Sanjay

PB - Springer Verlag

T2 - 6th International Workshop on Abdominal Imaging: Computational and Clinical Applications, ABDI 2014 held in conjunction with 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014

Y2 - 14 September 2014 through 14 September 2014

ER -

Mahapatra D, Schüffler PJ, Tielbeek JAW, Puylaert CAJ, Makanyanga JC, Menys A et al. Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation. In Yoshida H, Näppi JJ, Saini S, editors, Abdominal Imaging: Computational and Clinical Applications - 6th International Workshop, ABDI 2014 held in conjunction with MICCAI 2014. Springer Verlag. 2014. p. 139-147. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-13692-9_13

Combining multiple expert annotations using semi-supervised learning and graph cuts for Crohn’s disease segmentation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this