Confidence-guided sequential label fusion for multi-atlas based segmentation

Daoqiang Zhang, Guorong Wu, Hongjun Jia, Dinggang Shen

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

Abstract

Label fusion is a key step in multi-atlas based segmentation, which combines labels from multiple atlases to make the final decision. However, most of the current label fusion methods consider each voxel equally and independently during label fusion. In our point of view, however, different voxels act different roles in the way that some voxels might have much higher confidence in label determination than others, i.e., because of their better alignment across all registered atlases. In light of this, we propose a sequential label fusion framework for multi-atlas based image segmentation by hierarchically using the voxels with high confidence to guide the labeling procedure of other challenging voxels (whose registration results among deformed atlases are not good enough) to afford more accurate label fusion. Specifically, we first measure the corresponding labeling confidence for each voxel based on the k-nearest-neighbor rule, and then perform label fusion sequentially according to the estimated labeling confidence on each voxel. In particular, for each label fusion process, we use not only the propagated labels from atlases, but also the estimated labels from the neighboring voxels with higher labeling confidence. We demonstrate the advantage of our method by deploying it to the two popular label fusion algorithms, i.e., majority voting and local weighted voting. Experimental results show that our sequential label fusion method can consistently improve the performance of both algorithms in terms of segmentation/labeling accuracy.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages643-650
Number of pages8
Volume6893 LNCS
EditionPART 3
DOIs
Publication statusPublished - 2011 Oct 11
Externally publishedYes
Event14th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2011 - Toronto, ON, Canada
Duration: 2011 Sep 182011 Sep 22

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 3
Volume6893 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other14th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2011
CountryCanada
CityToronto, ON
Period11/9/1811/9/22

Fingerprint

Atlas
Voxel
Confidence
Labels
Fusion
Fusion reactions
Segmentation
Labeling
Majority Voting
Voting
Image Segmentation
Registration
Nearest Neighbor
Alignment
Image segmentation
Experimental Results

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Zhang, D., Wu, G., Jia, H., & Shen, D. (2011). Confidence-guided sequential label fusion for multi-atlas based segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 3 ed., Vol. 6893 LNCS, pp. 643-650). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6893 LNCS, No. PART 3). https://doi.org/10.1007/978-3-642-23626-6_79

Confidence-guided sequential label fusion for multi-atlas based segmentation. / Zhang, Daoqiang; Wu, Guorong; Jia, Hongjun; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6893 LNCS PART 3. ed. 2011. p. 643-650 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6893 LNCS, No. PART 3).

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

Zhang, D, Wu, G, Jia, H & Shen, D 2011, Confidence-guided sequential label fusion for multi-atlas based segmentation. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 3 edn, vol. 6893 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 3, vol. 6893 LNCS, pp. 643-650, 14th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2011, Toronto, ON, Canada, 11/9/18. https://doi.org/10.1007/978-3-642-23626-6_79
Zhang D, Wu G, Jia H, Shen D. Confidence-guided sequential label fusion for multi-atlas based segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 3 ed. Vol. 6893 LNCS. 2011. p. 643-650. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3). https://doi.org/10.1007/978-3-642-23626-6_79
Zhang, Daoqiang ; Wu, Guorong ; Jia, Hongjun ; Shen, Dinggang. / Confidence-guided sequential label fusion for multi-atlas based segmentation. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6893 LNCS PART 3. ed. 2011. pp. 643-650 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3).
@inproceedings{fe58b5ea15f24b08bda73aecb1e9d108,
title = "Confidence-guided sequential label fusion for multi-atlas based segmentation",
abstract = "Label fusion is a key step in multi-atlas based segmentation, which combines labels from multiple atlases to make the final decision. However, most of the current label fusion methods consider each voxel equally and independently during label fusion. In our point of view, however, different voxels act different roles in the way that some voxels might have much higher confidence in label determination than others, i.e., because of their better alignment across all registered atlases. In light of this, we propose a sequential label fusion framework for multi-atlas based image segmentation by hierarchically using the voxels with high confidence to guide the labeling procedure of other challenging voxels (whose registration results among deformed atlases are not good enough) to afford more accurate label fusion. Specifically, we first measure the corresponding labeling confidence for each voxel based on the k-nearest-neighbor rule, and then perform label fusion sequentially according to the estimated labeling confidence on each voxel. In particular, for each label fusion process, we use not only the propagated labels from atlases, but also the estimated labels from the neighboring voxels with higher labeling confidence. We demonstrate the advantage of our method by deploying it to the two popular label fusion algorithms, i.e., majority voting and local weighted voting. Experimental results show that our sequential label fusion method can consistently improve the performance of both algorithms in terms of segmentation/labeling accuracy.",
author = "Daoqiang Zhang and Guorong Wu and Hongjun Jia and Dinggang Shen",
year = "2011",
month = "10",
day = "11",
doi = "10.1007/978-3-642-23626-6_79",
language = "English",
isbn = "9783642236259",
volume = "6893 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 3",
pages = "643--650",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 3",

}

TY - GEN

T1 - Confidence-guided sequential label fusion for multi-atlas based segmentation

AU - Zhang, Daoqiang

AU - Wu, Guorong

AU - Jia, Hongjun

AU - Shen, Dinggang

PY - 2011/10/11

Y1 - 2011/10/11

N2 - Label fusion is a key step in multi-atlas based segmentation, which combines labels from multiple atlases to make the final decision. However, most of the current label fusion methods consider each voxel equally and independently during label fusion. In our point of view, however, different voxels act different roles in the way that some voxels might have much higher confidence in label determination than others, i.e., because of their better alignment across all registered atlases. In light of this, we propose a sequential label fusion framework for multi-atlas based image segmentation by hierarchically using the voxels with high confidence to guide the labeling procedure of other challenging voxels (whose registration results among deformed atlases are not good enough) to afford more accurate label fusion. Specifically, we first measure the corresponding labeling confidence for each voxel based on the k-nearest-neighbor rule, and then perform label fusion sequentially according to the estimated labeling confidence on each voxel. In particular, for each label fusion process, we use not only the propagated labels from atlases, but also the estimated labels from the neighboring voxels with higher labeling confidence. We demonstrate the advantage of our method by deploying it to the two popular label fusion algorithms, i.e., majority voting and local weighted voting. Experimental results show that our sequential label fusion method can consistently improve the performance of both algorithms in terms of segmentation/labeling accuracy.

AB - Label fusion is a key step in multi-atlas based segmentation, which combines labels from multiple atlases to make the final decision. However, most of the current label fusion methods consider each voxel equally and independently during label fusion. In our point of view, however, different voxels act different roles in the way that some voxels might have much higher confidence in label determination than others, i.e., because of their better alignment across all registered atlases. In light of this, we propose a sequential label fusion framework for multi-atlas based image segmentation by hierarchically using the voxels with high confidence to guide the labeling procedure of other challenging voxels (whose registration results among deformed atlases are not good enough) to afford more accurate label fusion. Specifically, we first measure the corresponding labeling confidence for each voxel based on the k-nearest-neighbor rule, and then perform label fusion sequentially according to the estimated labeling confidence on each voxel. In particular, for each label fusion process, we use not only the propagated labels from atlases, but also the estimated labels from the neighboring voxels with higher labeling confidence. We demonstrate the advantage of our method by deploying it to the two popular label fusion algorithms, i.e., majority voting and local weighted voting. Experimental results show that our sequential label fusion method can consistently improve the performance of both algorithms in terms of segmentation/labeling accuracy.

UR - http://www.scopus.com/inward/record.url?scp=80053477970&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053477970&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-23626-6_79

DO - 10.1007/978-3-642-23626-6_79

M3 - Conference contribution

SN - 9783642236259

VL - 6893 LNCS

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

SP - 643

EP - 650

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

ER -