Label stability in multiple instance learning

Veronika Cheplygina; Lauge Sørensen; David M. J. Tax; Marleen de Bruijne; Marco Loog

doi:10.1007/978-3-319-24553-9_66

Label stability in multiple instance learning

Veronika Cheplygina, Lauge Sørensen, David M. J. Tax, Marleen de Bruijne, Marco Loog

10 Citationer (Scopus)

Abstract

We address the problem of instance label stability in multiple instance learning (MIL) classifiers. These classifiers are trained only on globally annotated images (bags), but often can provide fine-grained annotations for image pixels or patches (instances). This is interesting for computer aided diagnosis (CAD) and other medical image analysis tasks for which only a coarse labeling is provided. Unfortunately, the instance labels may be unstable. This means that a slight change in training data could potentially lead to abnormalities being detected in different parts of the image, which is undesirable from a CAD point of view. Despite MIL gaining popularity in the CAD literature, this issue has not yet been addressed. We investigate the stability of instance labels provided by several MIL classifiers on 5 different datasets, of which 3 are medical image datasets (breast histopathology, diabetic retinopathy and computed tomography lung images).We propose an unsupervised measure to evaluate instance stability, and demonstrate that a performance-stability trade-off can be made when comparing MIL classifiers.

Originalsprog	Engelsk
Titel	Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015 : 18th international conference, Munich, Germany, October 5-9, 2015, Proceedings, Part I
Redaktører	Nassir Navab, Joachim Hornegger, William M. Wells, Alejandro F. Frangi
Antal sider	8
Forlag	Springer
Publikationsdato	2015
Sider	539-546
ISBN (Trykt)	978-3-319-24552-2
ISBN (Elektronisk)	978-3-319-24553-9
DOI	https://doi.org/10.1007/978-3-319-24553-9_66
Status	Udgivet - 2015
Begivenhed	International Conference on Medical Image Computing and Computer Assisted Intervention 2015 - Munich, Tyskland Varighed: 5 okt. 2015 → 9 okt. 2015 Konferencens nummer: 18

Konference

Konference	International Conference on Medical Image Computing and Computer Assisted Intervention 2015
Nummer	18
Land/Område	Tyskland
By	Munich
Periode	05/10/2015 → 09/10/2015

Navn	Lecture notes in computer science
ISSN	0302-9743

FN’s Verdensmål

Dette resultat bidrager til følgende verdensmål

Adgang til dokumentet

10.1007/978-3-319-24553-9_66

https://link.springer.com/content/pdf/10.1007/978-3-319-24553-9_66.pdf

Citationsformater

Cheplygina, V., Sørensen, L., Tax, D. M. J., de Bruijne, M., & Loog, M. (2015). Label stability in multiple instance learning. I N. Navab, J. Hornegger, W. M. Wells, & A. F. Frangi (red.), Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, Proceedings, Part I (s. 539-546). Springer. Lecture notes in computer science https://doi.org/10.1007/978-3-319-24553-9_66

Label stability in multiple instance learning. / Cheplygina, Veronika; Sørensen, Lauge; Tax, David M. J. et al.

Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, Proceedings, Part I. red. / Nassir Navab; Joachim Hornegger; William M. Wells; Alejandro F. Frangi. Springer, 2015. s. 539-546 (Lecture notes in computer science).

Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › peer review

Cheplygina, V, Sørensen, L, Tax, DMJ, de Bruijne, M & Loog, M 2015, Label stability in multiple instance learning. i N Navab, J Hornegger, WM Wells & AF Frangi (red), Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, Proceedings, Part I. Springer, Lecture notes in computer science, s. 539-546, International Conference on Medical Image Computing and Computer Assisted Intervention 2015, Munich, Tyskland, 05/10/2015. https://doi.org/10.1007/978-3-319-24553-9_66

Cheplygina V, Sørensen L, Tax DMJ, de Bruijne M, Loog M. Label stability in multiple instance learning. I Navab N, Hornegger J, Wells WM, Frangi AF, red., Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, Proceedings, Part I. Springer. 2015. s. 539-546. (Lecture notes in computer science). doi: 10.1007/978-3-319-24553-9_66

Cheplygina, Veronika ; Sørensen, Lauge ; Tax, David M. J. et al. / Label stability in multiple instance learning. Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, Proceedings, Part I. red. / Nassir Navab ; Joachim Hornegger ; William M. Wells ; Alejandro F. Frangi. Springer, 2015. s. 539-546 (Lecture notes in computer science).

@inproceedings{f07d28595dc641088fe08cc9b6b73a3a,

title = "Label stability in multiple instance learning",

abstract = "We address the problem of instance label stability in multiple instance learning (MIL) classifiers. These classifiers are trained only on globally annotated images (bags), but often can provide fine-grained annotations for image pixels or patches (instances). This is interesting for computer aided diagnosis (CAD) and other medical image analysis tasks for which only a coarse labeling is provided. Unfortunately, the instance labels may be unstable. This means that a slight change in training data could potentially lead to abnormalities being detected in different parts of the image, which is undesirable from a CAD point of view. Despite MIL gaining popularity in the CAD literature, this issue has not yet been addressed. We investigate the stability of instance labels provided by several MIL classifiers on 5 different datasets, of which 3 are medical image datasets (breast histopathology, diabetic retinopathy and computed tomography lung images).We propose an unsupervised measure to evaluate instance stability, and demonstrate that a performance-stability trade-off can be made when comparing MIL classifiers.",

author = "Veronika Cheplygina and Lauge S{\o}rensen and Tax, {David M. J.} and {de Bruijne}, Marleen and Marco Loog",

year = "2015",

doi = "10.1007/978-3-319-24553-9_66",

language = "English",

isbn = "978-3-319-24552-2",

series = "Lecture notes in computer science",

publisher = "Springer",

pages = "539--546",

editor = "Nassir Navab and Joachim Hornegger and Wells, {William M.} and Frangi, {Alejandro F.}",

booktitle = "Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015",

note = "International Conference on Medical Image Computing and Computer Assisted Intervention 2015 ; Conference date: 05-10-2015 Through 09-10-2015",

}

TY - GEN

T1 - Label stability in multiple instance learning

AU - Cheplygina, Veronika

AU - Sørensen, Lauge

AU - Tax, David M. J.

AU - de Bruijne, Marleen

AU - Loog, Marco

N1 - Conference code: 18

PY - 2015

Y1 - 2015

N2 - We address the problem of instance label stability in multiple instance learning (MIL) classifiers. These classifiers are trained only on globally annotated images (bags), but often can provide fine-grained annotations for image pixels or patches (instances). This is interesting for computer aided diagnosis (CAD) and other medical image analysis tasks for which only a coarse labeling is provided. Unfortunately, the instance labels may be unstable. This means that a slight change in training data could potentially lead to abnormalities being detected in different parts of the image, which is undesirable from a CAD point of view. Despite MIL gaining popularity in the CAD literature, this issue has not yet been addressed. We investigate the stability of instance labels provided by several MIL classifiers on 5 different datasets, of which 3 are medical image datasets (breast histopathology, diabetic retinopathy and computed tomography lung images).We propose an unsupervised measure to evaluate instance stability, and demonstrate that a performance-stability trade-off can be made when comparing MIL classifiers.

AB - We address the problem of instance label stability in multiple instance learning (MIL) classifiers. These classifiers are trained only on globally annotated images (bags), but often can provide fine-grained annotations for image pixels or patches (instances). This is interesting for computer aided diagnosis (CAD) and other medical image analysis tasks for which only a coarse labeling is provided. Unfortunately, the instance labels may be unstable. This means that a slight change in training data could potentially lead to abnormalities being detected in different parts of the image, which is undesirable from a CAD point of view. Despite MIL gaining popularity in the CAD literature, this issue has not yet been addressed. We investigate the stability of instance labels provided by several MIL classifiers on 5 different datasets, of which 3 are medical image datasets (breast histopathology, diabetic retinopathy and computed tomography lung images).We propose an unsupervised measure to evaluate instance stability, and demonstrate that a performance-stability trade-off can be made when comparing MIL classifiers.

U2 - 10.1007/978-3-319-24553-9_66

DO - 10.1007/978-3-319-24553-9_66

M3 - Article in proceedings

SN - 978-3-319-24552-2

T3 - Lecture notes in computer science

SP - 539

EP - 546

BT - Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015

A2 - Navab, Nassir

A2 - Hornegger, Joachim

A2 - Wells, William M.

A2 - Frangi, Alejandro F.

PB - Springer

T2 - International Conference on Medical Image Computing and Computer Assisted Intervention 2015

Y2 - 5 October 2015 through 9 October 2015

ER -

Label stability in multiple instance learning

Abstract

Konference

FN’s Verdensmål

Adgang til dokumentet

Fingeraftryk

Citationsformater