Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels)

Lin Gu, Yinqiang Zheng, Ryoma Bise, Imari Sato, Nobuaki Imanishi, Sadakazu Aiso

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

8 Citations (Scopus)

Abstract

In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.

Original languageEnglish
Title of host publicationMedical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings
PublisherSpringer Verlag
Pages702-710
Number of pages9
Volume10433 LNCS
ISBN (Print)9783319661810
DOIs
Publication statusPublished - 2017
Event20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017 - Quebec City, Canada
Duration: 2017 Sep 112017 Sep 13

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10433 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017
CountryCanada
CityQuebec City
Period17/9/1117/9/13

Fingerprint

Semi-supervised Learning
Supervised learning
Voxel
Image segmentation
Image Segmentation
Classifiers
Pixel
Pixels
Classifier
Random Forest
Confidence
MATLAB
Segmentation
Tend

Keywords

  • Image segmentation
  • Random forest
  • Semi-supervised learning
  • Super pixels(voxels)

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Gu, L., Zheng, Y., Bise, R., Sato, I., Imanishi, N., & Aiso, S. (2017). Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). In Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings (Vol. 10433 LNCS, pp. 702-710). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10433 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-66182-7_80

Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). / Gu, Lin; Zheng, Yinqiang; Bise, Ryoma; Sato, Imari; Imanishi, Nobuaki; Aiso, Sadakazu.

Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Vol. 10433 LNCS Springer Verlag, 2017. p. 702-710 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10433 LNCS).

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

Gu, L, Zheng, Y, Bise, R, Sato, I, Imanishi, N & Aiso, S 2017, Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). in Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. vol. 10433 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10433 LNCS, Springer Verlag, pp. 702-710, 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017, Quebec City, Canada, 17/9/11. https://doi.org/10.1007/978-3-319-66182-7_80
Gu L, Zheng Y, Bise R, Sato I, Imanishi N, Aiso S. Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). In Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Vol. 10433 LNCS. Springer Verlag. 2017. p. 702-710. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-66182-7_80
Gu, Lin ; Zheng, Yinqiang ; Bise, Ryoma ; Sato, Imari ; Imanishi, Nobuaki ; Aiso, Sadakazu. / Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Vol. 10433 LNCS Springer Verlag, 2017. pp. 702-710 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{020e7c4542f14bdaa20a148a4b276066,
title = "Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels)",
abstract = "In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.",
keywords = "Image segmentation, Random forest, Semi-supervised learning, Super pixels(voxels)",
author = "Lin Gu and Yinqiang Zheng and Ryoma Bise and Imari Sato and Nobuaki Imanishi and Sadakazu Aiso",
year = "2017",
doi = "10.1007/978-3-319-66182-7_80",
language = "English",
isbn = "9783319661810",
volume = "10433 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "702--710",
booktitle = "Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings",
address = "Germany",

}

TY - GEN

T1 - Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels)

AU - Gu, Lin

AU - Zheng, Yinqiang

AU - Bise, Ryoma

AU - Sato, Imari

AU - Imanishi, Nobuaki

AU - Aiso, Sadakazu

PY - 2017

Y1 - 2017

N2 - In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.

AB - In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.

KW - Image segmentation

KW - Random forest

KW - Semi-supervised learning

KW - Super pixels(voxels)

UR - http://www.scopus.com/inward/record.url?scp=85029384171&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029384171&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-66182-7_80

DO - 10.1007/978-3-319-66182-7_80

M3 - Conference contribution

SN - 9783319661810

VL - 10433 LNCS

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

SP - 702

EP - 710

BT - Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings

PB - Springer Verlag

ER -