Multi-atlas based segmentation editing with interaction-guided constraints

Sang Hyun Park, Yaozong Gao, Dinggang Shen

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

1 Citation (Scopus)

Abstract

We propose a novel multi-atlas based segmentation method to address the editing scenario, when given an incomplete segmentation along with a set of training label images. Unlike previous multi-atlas based methods, which depend solely on appearance features, we incorporate interaction-guided constraints to find appropriate training labels and derive their voting weights. Specifically, we divide user interactions, provided on erroneous parts, into multiple local interaction combinations, and then locally search for the training label patches well-matched with each interaction combination and also the previous segmentation. Then, we estimate the new segmentation through the label fusion of selected label patches that have their weights defined with respect to their respective distances to the interactions. Since the label patches are found to be from different combinations in our method, various shape changes can be considered even with limited training labels and few user interactions. Since our method does not need image information or expensive learning steps, it can be conveniently used for most editing problems. To demonstrate the positive performance, we apply our method to editing the segmentation of three challenging data sets: prostate CT, brainstem CT, and hippocampus MR. The results show that our method outperforms the existing editing methods in all three data sets.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages198-206
Number of pages9
Volume9351
ISBN (Print)9783319245737
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015 - Munich, Germany
Duration: 2015 Oct 52015 Oct 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9351
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015
CountryGermany
CityMunich
Period15/10/515/10/9

Fingerprint

Atlas
Labels
Segmentation
Interaction
Patch
User Interaction
Hippocampus
Local Interaction
Voting
Divides
Fusion
Fusion reactions
Scenarios
Training
Estimate
Demonstrate

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Park, S. H., Gao, Y., & Shen, D. (2015). Multi-atlas based segmentation editing with interaction-guided constraints. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9351, pp. 198-206). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351). Springer Verlag. https://doi.org/10.1007/978-3-319-24574-4_24

Multi-atlas based segmentation editing with interaction-guided constraints. / Park, Sang Hyun; Gao, Yaozong; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. p. 198-206 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351).

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

Park, SH, Gao, Y & Shen, D 2015, Multi-atlas based segmentation editing with interaction-guided constraints. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9351, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9351, Springer Verlag, pp. 198-206, 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015, Munich, Germany, 15/10/5. https://doi.org/10.1007/978-3-319-24574-4_24
Park SH, Gao Y, Shen D. Multi-atlas based segmentation editing with interaction-guided constraints. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351. Springer Verlag. 2015. p. 198-206. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-24574-4_24
Park, Sang Hyun ; Gao, Yaozong ; Shen, Dinggang. / Multi-atlas based segmentation editing with interaction-guided constraints. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. pp. 198-206 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{f6e821c1732241b4a54ffe015bd76184,
title = "Multi-atlas based segmentation editing with interaction-guided constraints",
abstract = "We propose a novel multi-atlas based segmentation method to address the editing scenario, when given an incomplete segmentation along with a set of training label images. Unlike previous multi-atlas based methods, which depend solely on appearance features, we incorporate interaction-guided constraints to find appropriate training labels and derive their voting weights. Specifically, we divide user interactions, provided on erroneous parts, into multiple local interaction combinations, and then locally search for the training label patches well-matched with each interaction combination and also the previous segmentation. Then, we estimate the new segmentation through the label fusion of selected label patches that have their weights defined with respect to their respective distances to the interactions. Since the label patches are found to be from different combinations in our method, various shape changes can be considered even with limited training labels and few user interactions. Since our method does not need image information or expensive learning steps, it can be conveniently used for most editing problems. To demonstrate the positive performance, we apply our method to editing the segmentation of three challenging data sets: prostate CT, brainstem CT, and hippocampus MR. The results show that our method outperforms the existing editing methods in all three data sets.",
author = "Park, {Sang Hyun} and Yaozong Gao and Dinggang Shen",
year = "2015",
doi = "10.1007/978-3-319-24574-4_24",
language = "English",
isbn = "9783319245737",
volume = "9351",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "198--206",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Multi-atlas based segmentation editing with interaction-guided constraints

AU - Park, Sang Hyun

AU - Gao, Yaozong

AU - Shen, Dinggang

PY - 2015

Y1 - 2015

N2 - We propose a novel multi-atlas based segmentation method to address the editing scenario, when given an incomplete segmentation along with a set of training label images. Unlike previous multi-atlas based methods, which depend solely on appearance features, we incorporate interaction-guided constraints to find appropriate training labels and derive their voting weights. Specifically, we divide user interactions, provided on erroneous parts, into multiple local interaction combinations, and then locally search for the training label patches well-matched with each interaction combination and also the previous segmentation. Then, we estimate the new segmentation through the label fusion of selected label patches that have their weights defined with respect to their respective distances to the interactions. Since the label patches are found to be from different combinations in our method, various shape changes can be considered even with limited training labels and few user interactions. Since our method does not need image information or expensive learning steps, it can be conveniently used for most editing problems. To demonstrate the positive performance, we apply our method to editing the segmentation of three challenging data sets: prostate CT, brainstem CT, and hippocampus MR. The results show that our method outperforms the existing editing methods in all three data sets.

AB - We propose a novel multi-atlas based segmentation method to address the editing scenario, when given an incomplete segmentation along with a set of training label images. Unlike previous multi-atlas based methods, which depend solely on appearance features, we incorporate interaction-guided constraints to find appropriate training labels and derive their voting weights. Specifically, we divide user interactions, provided on erroneous parts, into multiple local interaction combinations, and then locally search for the training label patches well-matched with each interaction combination and also the previous segmentation. Then, we estimate the new segmentation through the label fusion of selected label patches that have their weights defined with respect to their respective distances to the interactions. Since the label patches are found to be from different combinations in our method, various shape changes can be considered even with limited training labels and few user interactions. Since our method does not need image information or expensive learning steps, it can be conveniently used for most editing problems. To demonstrate the positive performance, we apply our method to editing the segmentation of three challenging data sets: prostate CT, brainstem CT, and hippocampus MR. The results show that our method outperforms the existing editing methods in all three data sets.

UR - http://www.scopus.com/inward/record.url?scp=84951834251&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951834251&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-24574-4_24

DO - 10.1007/978-3-319-24574-4_24

M3 - Conference contribution

AN - SCOPUS:84951834251

SN - 9783319245737

VL - 9351

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

SP - 198

EP - 206

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

PB - Springer Verlag

ER -