Exploring the similarity of medical imaging classification problems

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)

Abstract

Supervised learning is ubiquitous in medical image analysis. In this paper we consider the problem of meta-learning – predicting which methods will perform well in an unseen classification problem, given previous experience with other classification problems. We investigate the first step of such an approach: how to quantify the similarity of different classification problems. We characterize datasets sampled from six classification problems by performance ranks of simple classifiers, and define the similarity by the inverse of Euclidean distance in this meta-feature space. We visualize the similarities in a 2D space, where meaningful clusters start to emerge, and show that the proposed representation can be used to classify datasets according to their origin with 89.3% accuracy. These findings, together with the observations of recent trends in machine learning, suggest that meta-learning could be a valuable tool for the medical imaging community.

LanguageEnglish
Title of host publicationIntravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis
Subtitle of host publication 6th Joint International Workshops, CVII-STENT 2017 and 2nd International Workshop, LABELS 2017 Held in Conjunction with MICCAI 2017, Proceedings
EditorsM.J. Cardoso, T. Arbel, et. al.
Place of PublicationCham
PublisherSpringer
Pages59-66
Number of pages8
ISBN (Electronic)978-3-319-67534-3
ISBN (Print)9783319675336
DOIs
StatePublished - 2017
Event2nd International Workshop on Large-Scale Annotation of Biomedical Data and Expert Label Synthesis, (LABELS 2017), 10-14 September 2017, Quebec City, Canada - Quebec City, Canada
Duration: 10 Sep 201714 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10552 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Workshop on Large-Scale Annotation of Biomedical Data and Expert Label Synthesis, (LABELS 2017), 10-14 September 2017, Quebec City, Canada
CountryCanada
CityQuebec City
Period10/09/1714/09/17
OtherHeld in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017

Fingerprint

Medical Imaging
Medical imaging
Classification Problems
Meta-learning
Medical Image Analysis
Supervised learning
Supervised Learning
Euclidean Distance
Feature Space
Image analysis
Learning systems
Machine Learning
Quantify
Classifiers
Classify
Classifier
Similarity

Cite this

Cheplygina, V., Moeskops, P., Veta, M., Dashtbozorg, B., & Pluim, J. P. W. (2017). Exploring the similarity of medical imaging classification problems. In M. J. Cardoso, T. Arbel, & et. al. (Eds.), Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis: 6th Joint International Workshops, CVII-STENT 2017 and 2nd International Workshop, LABELS 2017 Held in Conjunction with MICCAI 2017, Proceedings (pp. 59-66). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10552 LNCS). Cham: Springer. DOI: 10.1007/978-3-319-67534-3_7
Cheplygina, V. ; Moeskops, P. ; Veta, M. ; Dashtbozorg, B. ; Pluim, J.P.W./ Exploring the similarity of medical imaging classification problems. Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis: 6th Joint International Workshops, CVII-STENT 2017 and 2nd International Workshop, LABELS 2017 Held in Conjunction with MICCAI 2017, Proceedings. editor / M.J. Cardoso ; T. Arbel ; et. al.Cham : Springer, 2017. pp. 59-66 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{62d7edb2c6734601aca4c53b90a6f3c2,
title = "Exploring the similarity of medical imaging classification problems",
abstract = "Supervised learning is ubiquitous in medical image analysis. In this paper we consider the problem of meta-learning – predicting which methods will perform well in an unseen classification problem, given previous experience with other classification problems. We investigate the first step of such an approach: how to quantify the similarity of different classification problems. We characterize datasets sampled from six classification problems by performance ranks of simple classifiers, and define the similarity by the inverse of Euclidean distance in this meta-feature space. We visualize the similarities in a 2D space, where meaningful clusters start to emerge, and show that the proposed representation can be used to classify datasets according to their origin with 89.3{\%} accuracy. These findings, together with the observations of recent trends in machine learning, suggest that meta-learning could be a valuable tool for the medical imaging community.",
author = "V. Cheplygina and P. Moeskops and M. Veta and B. Dashtbozorg and J.P.W. Pluim",
year = "2017",
doi = "10.1007/978-3-319-67534-3_7",
language = "English",
isbn = "9783319675336",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "59--66",
editor = "M.J. Cardoso and T. Arbel and {et. al.}",
booktitle = "Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis",
address = "Germany",

}

Cheplygina, V, Moeskops, P, Veta, M, Dashtbozorg, B & Pluim, JPW 2017, Exploring the similarity of medical imaging classification problems. in MJ Cardoso, T Arbel & et. al. (eds), Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis: 6th Joint International Workshops, CVII-STENT 2017 and 2nd International Workshop, LABELS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10552 LNCS, Springer, Cham, pp. 59-66, 2nd International Workshop on Large-Scale Annotation of Biomedical Data and Expert Label Synthesis, (LABELS 2017), 10-14 September 2017, Quebec City, Canada, Quebec City, Canada, 10/09/17. DOI: 10.1007/978-3-319-67534-3_7

Exploring the similarity of medical imaging classification problems. / Cheplygina, V.; Moeskops, P.; Veta, M.; Dashtbozorg, B.; Pluim, J.P.W.

Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis: 6th Joint International Workshops, CVII-STENT 2017 and 2nd International Workshop, LABELS 2017 Held in Conjunction with MICCAI 2017, Proceedings. ed. / M.J. Cardoso; T. Arbel; et. al.Cham : Springer, 2017. p. 59-66 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10552 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Exploring the similarity of medical imaging classification problems

AU - Cheplygina,V.

AU - Moeskops,P.

AU - Veta,M.

AU - Dashtbozorg,B.

AU - Pluim,J.P.W.

PY - 2017

Y1 - 2017

N2 - Supervised learning is ubiquitous in medical image analysis. In this paper we consider the problem of meta-learning – predicting which methods will perform well in an unseen classification problem, given previous experience with other classification problems. We investigate the first step of such an approach: how to quantify the similarity of different classification problems. We characterize datasets sampled from six classification problems by performance ranks of simple classifiers, and define the similarity by the inverse of Euclidean distance in this meta-feature space. We visualize the similarities in a 2D space, where meaningful clusters start to emerge, and show that the proposed representation can be used to classify datasets according to their origin with 89.3% accuracy. These findings, together with the observations of recent trends in machine learning, suggest that meta-learning could be a valuable tool for the medical imaging community.

AB - Supervised learning is ubiquitous in medical image analysis. In this paper we consider the problem of meta-learning – predicting which methods will perform well in an unseen classification problem, given previous experience with other classification problems. We investigate the first step of such an approach: how to quantify the similarity of different classification problems. We characterize datasets sampled from six classification problems by performance ranks of simple classifiers, and define the similarity by the inverse of Euclidean distance in this meta-feature space. We visualize the similarities in a 2D space, where meaningful clusters start to emerge, and show that the proposed representation can be used to classify datasets according to their origin with 89.3% accuracy. These findings, together with the observations of recent trends in machine learning, suggest that meta-learning could be a valuable tool for the medical imaging community.

UR - http://www.scopus.com/inward/record.url?scp=85029820846&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-67534-3_7

DO - 10.1007/978-3-319-67534-3_7

M3 - Conference contribution

SN - 9783319675336

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

SP - 59

EP - 66

BT - Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis

PB - Springer

CY - Cham

ER -

Cheplygina V, Moeskops P, Veta M, Dashtbozorg B, Pluim JPW. Exploring the similarity of medical imaging classification problems. In Cardoso MJ, Arbel T, et. al., editors, Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis: 6th Joint International Workshops, CVII-STENT 2017 and 2nd International Workshop, LABELS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Cham: Springer. 2017. p. 59-66. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Available from, DOI: 10.1007/978-3-319-67534-3_7