Dice overlap measures for objects of unknown number

Application to lesion segmentation

Ipek Oguz, Aaron Carass, Dzung L. Pham, Snehashis Roy, Nagesh Subbana, Peter Calabresi, Paul A. Yushkevich, Russell T. Shinohara, Jerry Ladd Prince

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

Abstract

The Dice overlap ratio is commonly used to evaluate the performance of image segmentation algorithms. While Dice overlap is very useful as a standardized quantitative measure of segmentation accuracy in many applications, it offers a very limited picture of segmentation quality in complex segmentation tasks where the number of target objects is not known a priori, such as the segmentation of white matter lesions or lung nodules. While Dice overlap can still be used in these applications, segmentation algorithms may perform quite differently in ways not reflected by differences in their Dice score. Here we propose a new set of evaluation techniques that offer new insights into the behavior of segmentation algorithms. We illustrate these techniques with a case study comparing two popular multiple sclerosis (MS) lesion segmentation algorithms: OASIS and LesionTOADS.

Original languageEnglish (US)
Title of host publicationBrainlesion
Subtitle of host publicationGlioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers
PublisherSpringer Verlag
Pages3-14
Number of pages12
ISBN (Print)9783319752372
DOIs
StatePublished - Jan 1 2018
Event3rd International Workshop on Brainlesion, BrainLes 2017 Held in Conjunction with Medical Image Computing for Computer Assisted Intervention , MICCAI 2017 - Quebec City, Canada
Duration: Sep 14 2017Sep 14 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10670 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other3rd International Workshop on Brainlesion, BrainLes 2017 Held in Conjunction with Medical Image Computing for Computer Assisted Intervention , MICCAI 2017
CountryCanada
CityQuebec City
Period9/14/179/14/17

Fingerprint

Dice
Overlap
Segmentation
Unknown
Image segmentation
Multiple Sclerosis
Nodule
Object
Lung
Image Segmentation
Target
Evaluate
Evaluation

Keywords

  • Evaluation
  • Lesion
  • MS
  • Segmentation

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Oguz, I., Carass, A., Pham, D. L., Roy, S., Subbana, N., Calabresi, P., ... Prince, J. L. (2018). Dice overlap measures for objects of unknown number: Application to lesion segmentation. In Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers (pp. 3-14). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10670 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-75238-9_1

Dice overlap measures for objects of unknown number : Application to lesion segmentation. / Oguz, Ipek; Carass, Aaron; Pham, Dzung L.; Roy, Snehashis; Subbana, Nagesh; Calabresi, Peter; Yushkevich, Paul A.; Shinohara, Russell T.; Prince, Jerry Ladd.

Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Springer Verlag, 2018. p. 3-14 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10670 LNCS).

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

Oguz, I, Carass, A, Pham, DL, Roy, S, Subbana, N, Calabresi, P, Yushkevich, PA, Shinohara, RT & Prince, JL 2018, Dice overlap measures for objects of unknown number: Application to lesion segmentation. in Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10670 LNCS, Springer Verlag, pp. 3-14, 3rd International Workshop on Brainlesion, BrainLes 2017 Held in Conjunction with Medical Image Computing for Computer Assisted Intervention , MICCAI 2017, Quebec City, Canada, 9/14/17. https://doi.org/10.1007/978-3-319-75238-9_1
Oguz I, Carass A, Pham DL, Roy S, Subbana N, Calabresi P et al. Dice overlap measures for objects of unknown number: Application to lesion segmentation. In Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Springer Verlag. 2018. p. 3-14. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-75238-9_1
Oguz, Ipek ; Carass, Aaron ; Pham, Dzung L. ; Roy, Snehashis ; Subbana, Nagesh ; Calabresi, Peter ; Yushkevich, Paul A. ; Shinohara, Russell T. ; Prince, Jerry Ladd. / Dice overlap measures for objects of unknown number : Application to lesion segmentation. Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Springer Verlag, 2018. pp. 3-14 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c131f37fed2e4bf28c1f3f590fcfbd44,
title = "Dice overlap measures for objects of unknown number: Application to lesion segmentation",
abstract = "The Dice overlap ratio is commonly used to evaluate the performance of image segmentation algorithms. While Dice overlap is very useful as a standardized quantitative measure of segmentation accuracy in many applications, it offers a very limited picture of segmentation quality in complex segmentation tasks where the number of target objects is not known a priori, such as the segmentation of white matter lesions or lung nodules. While Dice overlap can still be used in these applications, segmentation algorithms may perform quite differently in ways not reflected by differences in their Dice score. Here we propose a new set of evaluation techniques that offer new insights into the behavior of segmentation algorithms. We illustrate these techniques with a case study comparing two popular multiple sclerosis (MS) lesion segmentation algorithms: OASIS and LesionTOADS.",
keywords = "Evaluation, Lesion, MS, Segmentation",
author = "Ipek Oguz and Aaron Carass and Pham, {Dzung L.} and Snehashis Roy and Nagesh Subbana and Peter Calabresi and Yushkevich, {Paul A.} and Shinohara, {Russell T.} and Prince, {Jerry Ladd}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-75238-9_1",
language = "English (US)",
isbn = "9783319752372",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "3--14",
booktitle = "Brainlesion",

}

TY - GEN

T1 - Dice overlap measures for objects of unknown number

T2 - Application to lesion segmentation

AU - Oguz, Ipek

AU - Carass, Aaron

AU - Pham, Dzung L.

AU - Roy, Snehashis

AU - Subbana, Nagesh

AU - Calabresi, Peter

AU - Yushkevich, Paul A.

AU - Shinohara, Russell T.

AU - Prince, Jerry Ladd

PY - 2018/1/1

Y1 - 2018/1/1

N2 - The Dice overlap ratio is commonly used to evaluate the performance of image segmentation algorithms. While Dice overlap is very useful as a standardized quantitative measure of segmentation accuracy in many applications, it offers a very limited picture of segmentation quality in complex segmentation tasks where the number of target objects is not known a priori, such as the segmentation of white matter lesions or lung nodules. While Dice overlap can still be used in these applications, segmentation algorithms may perform quite differently in ways not reflected by differences in their Dice score. Here we propose a new set of evaluation techniques that offer new insights into the behavior of segmentation algorithms. We illustrate these techniques with a case study comparing two popular multiple sclerosis (MS) lesion segmentation algorithms: OASIS and LesionTOADS.

AB - The Dice overlap ratio is commonly used to evaluate the performance of image segmentation algorithms. While Dice overlap is very useful as a standardized quantitative measure of segmentation accuracy in many applications, it offers a very limited picture of segmentation quality in complex segmentation tasks where the number of target objects is not known a priori, such as the segmentation of white matter lesions or lung nodules. While Dice overlap can still be used in these applications, segmentation algorithms may perform quite differently in ways not reflected by differences in their Dice score. Here we propose a new set of evaluation techniques that offer new insights into the behavior of segmentation algorithms. We illustrate these techniques with a case study comparing two popular multiple sclerosis (MS) lesion segmentation algorithms: OASIS and LesionTOADS.

KW - Evaluation

KW - Lesion

KW - MS

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=85042530721&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042530721&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-75238-9_1

DO - 10.1007/978-3-319-75238-9_1

M3 - Conference contribution

SN - 9783319752372

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

SP - 3

EP - 14

BT - Brainlesion

PB - Springer Verlag

ER -