Learning-based atlas selection for multiple-atlas segmentation

Gerard Sanroma, Guorong Wu, Yaozong Gao, Dinggang Shen

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

12 Citations (Scopus)

Abstract

Recently, multi-atlas segmentation (MAS) has achieved a great success in the medical imaging area. The key assumption of MAS is that multiple atlases encompass richer anatomical variability than a single atlas. Therefore, we can label the target image more accurately by mapping the label information from the appropriate atlas images that have the most similar structures. The problem of atlas selection, however, still remains unexplored. Current state-of-the-art MAS methods rely on image similarity to select a set of atlases. Unfortunately, this heuristic criterion is not necessarily related to segmentation performance and, thus may undermine segmentation results. To solve this simple but critical problem, we propose a learning-based atlas selection method to pick up the best atlases that would eventually lead to more accurate image segmentation. Our idea is to learn the relationship between the pairwise appearance of observed instances (a pair of atlas and target images) and their final labeling performance (in terms of Dice ratio). In this way, we can select the best atlases according to their expected labeling accuracy. It is worth noting that our atlas selection method is general enough to be integrated with existing MAS methods. As is shown in the experiments, we achieve significant improvement after we integrate our method with 3 widely used MAS methods on ADNI and LONI LPBA40 datasets.

Original languageEnglish
Title of host publicationProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
PublisherIEEE Computer Society
Pages3111-3117
Number of pages7
ISBN (Print)9781479951178, 9781479951178
DOIs
Publication statusPublished - 2014 Jan 1
Externally publishedYes
Event27th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2014 - Columbus, United States
Duration: 2014 Jun 232014 Jun 28

Other

Other27th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2014
CountryUnited States
CityColumbus
Period14/6/2314/6/28

Fingerprint

Labeling
Labels
Medical imaging
Image segmentation
Experiments

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

Sanroma, G., Wu, G., Gao, Y., & Shen, D. (2014). Learning-based atlas selection for multiple-atlas segmentation. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (pp. 3111-3117). [6909794] IEEE Computer Society. https://doi.org/10.1109/CVPR.2014.398

Learning-based atlas selection for multiple-atlas segmentation. / Sanroma, Gerard; Wu, Guorong; Gao, Yaozong; Shen, Dinggang.

Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2014. p. 3111-3117 6909794.

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

Sanroma, G, Wu, G, Gao, Y & Shen, D 2014, Learning-based atlas selection for multiple-atlas segmentation. in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition., 6909794, IEEE Computer Society, pp. 3111-3117, 27th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2014, Columbus, United States, 14/6/23. https://doi.org/10.1109/CVPR.2014.398
Sanroma G, Wu G, Gao Y, Shen D. Learning-based atlas selection for multiple-atlas segmentation. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE Computer Society. 2014. p. 3111-3117. 6909794 https://doi.org/10.1109/CVPR.2014.398
Sanroma, Gerard ; Wu, Guorong ; Gao, Yaozong ; Shen, Dinggang. / Learning-based atlas selection for multiple-atlas segmentation. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2014. pp. 3111-3117
@inproceedings{21d92d4fa5654c838fc5626a076d706c,
title = "Learning-based atlas selection for multiple-atlas segmentation",
abstract = "Recently, multi-atlas segmentation (MAS) has achieved a great success in the medical imaging area. The key assumption of MAS is that multiple atlases encompass richer anatomical variability than a single atlas. Therefore, we can label the target image more accurately by mapping the label information from the appropriate atlas images that have the most similar structures. The problem of atlas selection, however, still remains unexplored. Current state-of-the-art MAS methods rely on image similarity to select a set of atlases. Unfortunately, this heuristic criterion is not necessarily related to segmentation performance and, thus may undermine segmentation results. To solve this simple but critical problem, we propose a learning-based atlas selection method to pick up the best atlases that would eventually lead to more accurate image segmentation. Our idea is to learn the relationship between the pairwise appearance of observed instances (a pair of atlas and target images) and their final labeling performance (in terms of Dice ratio). In this way, we can select the best atlases according to their expected labeling accuracy. It is worth noting that our atlas selection method is general enough to be integrated with existing MAS methods. As is shown in the experiments, we achieve significant improvement after we integrate our method with 3 widely used MAS methods on ADNI and LONI LPBA40 datasets.",
keywords = "Atlas selection, feature selection, multi-atlas based segmentation, SVM rank",
author = "Gerard Sanroma and Guorong Wu and Yaozong Gao and Dinggang Shen",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/CVPR.2014.398",
language = "English",
isbn = "9781479951178",
pages = "3111--3117",
booktitle = "Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Learning-based atlas selection for multiple-atlas segmentation

AU - Sanroma, Gerard

AU - Wu, Guorong

AU - Gao, Yaozong

AU - Shen, Dinggang

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Recently, multi-atlas segmentation (MAS) has achieved a great success in the medical imaging area. The key assumption of MAS is that multiple atlases encompass richer anatomical variability than a single atlas. Therefore, we can label the target image more accurately by mapping the label information from the appropriate atlas images that have the most similar structures. The problem of atlas selection, however, still remains unexplored. Current state-of-the-art MAS methods rely on image similarity to select a set of atlases. Unfortunately, this heuristic criterion is not necessarily related to segmentation performance and, thus may undermine segmentation results. To solve this simple but critical problem, we propose a learning-based atlas selection method to pick up the best atlases that would eventually lead to more accurate image segmentation. Our idea is to learn the relationship between the pairwise appearance of observed instances (a pair of atlas and target images) and their final labeling performance (in terms of Dice ratio). In this way, we can select the best atlases according to their expected labeling accuracy. It is worth noting that our atlas selection method is general enough to be integrated with existing MAS methods. As is shown in the experiments, we achieve significant improvement after we integrate our method with 3 widely used MAS methods on ADNI and LONI LPBA40 datasets.

AB - Recently, multi-atlas segmentation (MAS) has achieved a great success in the medical imaging area. The key assumption of MAS is that multiple atlases encompass richer anatomical variability than a single atlas. Therefore, we can label the target image more accurately by mapping the label information from the appropriate atlas images that have the most similar structures. The problem of atlas selection, however, still remains unexplored. Current state-of-the-art MAS methods rely on image similarity to select a set of atlases. Unfortunately, this heuristic criterion is not necessarily related to segmentation performance and, thus may undermine segmentation results. To solve this simple but critical problem, we propose a learning-based atlas selection method to pick up the best atlases that would eventually lead to more accurate image segmentation. Our idea is to learn the relationship between the pairwise appearance of observed instances (a pair of atlas and target images) and their final labeling performance (in terms of Dice ratio). In this way, we can select the best atlases according to their expected labeling accuracy. It is worth noting that our atlas selection method is general enough to be integrated with existing MAS methods. As is shown in the experiments, we achieve significant improvement after we integrate our method with 3 widely used MAS methods on ADNI and LONI LPBA40 datasets.

KW - Atlas selection

KW - feature selection

KW - multi-atlas based segmentation

KW - SVM rank

UR - http://www.scopus.com/inward/record.url?scp=84911393068&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84911393068&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2014.398

DO - 10.1109/CVPR.2014.398

M3 - Conference contribution

SN - 9781479951178

SN - 9781479951178

SP - 3111

EP - 3117

BT - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

PB - IEEE Computer Society

ER -