A fixed-point model for pancreas segmentation in abdominal CT scans

Yuyin Zhou, Lingxi Xie, Wei Shen, Yan Wang, Elliot K Fishman, Alan L. Yuille

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

Abstract

Deep neural networks have been widely adopted for automatic organ segmentation from abdominal CT scans. However, the segmentation accuracy of some small organs (e.g., the pancreas) is sometimes below satisfaction, arguably because deep networks are easily disrupted by the complex and variable background regions which occupies a large fraction of the input volume. In this paper, we formulate this problem into a fixed-point model which uses a predicted segmentation mask to shrink the input region. This is motivated by the fact that a smaller input region often leads to more accurate segmentation. In the training process, we use the ground-truth annotation to generate accurate input regions and optimize network weights. On the testing stage, we fix the network parameters and update the segmentation results in an iterative manner. We evaluate our approach on the NIH pancreas segmentation dataset, and outperform the state-of-the-art by more than 4%, measured by the average Dice-Sørensen Coefficient (DSC). In addition, we report 62.43% DSC in the worst case, which guarantees the reliability of our approach in clinical applications.

Original languageEnglish (US)
Title of host publicationMedical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings
PublisherSpringer Verlag
Pages693-701
Number of pages9
Volume10433 LNCS
ISBN (Print)9783319661810
DOIs
StatePublished - 2017
Event20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017 - Quebec City, Canada
Duration: Sep 11 2017Sep 13 2017

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
Period9/11/179/13/17

Fingerprint

Computerized tomography
Masks
Segmentation
Fixed point
Testing
Dice
Model
Coefficient
Mask
Annotation
Deep neural networks
Update
Optimise
Neural Networks
Evaluate

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Zhou, Y., Xie, L., Shen, W., Wang, Y., Fishman, E. K., & Yuille, A. L. (2017). A fixed-point model for pancreas segmentation in abdominal CT scans. In Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings (Vol. 10433 LNCS, pp. 693-701). (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_79

A fixed-point model for pancreas segmentation in abdominal CT scans. / Zhou, Yuyin; Xie, Lingxi; Shen, Wei; Wang, Yan; Fishman, Elliot K; Yuille, Alan L.

Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Vol. 10433 LNCS Springer Verlag, 2017. p. 693-701 (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

Zhou, Y, Xie, L, Shen, W, Wang, Y, Fishman, EK & Yuille, AL 2017, A fixed-point model for pancreas segmentation in abdominal CT scans. 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. 693-701, 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017, Quebec City, Canada, 9/11/17. https://doi.org/10.1007/978-3-319-66182-7_79
Zhou Y, Xie L, Shen W, Wang Y, Fishman EK, Yuille AL. A fixed-point model for pancreas segmentation in abdominal CT scans. In Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Vol. 10433 LNCS. Springer Verlag. 2017. p. 693-701. (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_79
Zhou, Yuyin ; Xie, Lingxi ; Shen, Wei ; Wang, Yan ; Fishman, Elliot K ; Yuille, Alan L. / A fixed-point model for pancreas segmentation in abdominal CT scans. Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Vol. 10433 LNCS Springer Verlag, 2017. pp. 693-701 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{85faba41de09418295dab4ce9d65568e,
title = "A fixed-point model for pancreas segmentation in abdominal CT scans",
abstract = "Deep neural networks have been widely adopted for automatic organ segmentation from abdominal CT scans. However, the segmentation accuracy of some small organs (e.g., the pancreas) is sometimes below satisfaction, arguably because deep networks are easily disrupted by the complex and variable background regions which occupies a large fraction of the input volume. In this paper, we formulate this problem into a fixed-point model which uses a predicted segmentation mask to shrink the input region. This is motivated by the fact that a smaller input region often leads to more accurate segmentation. In the training process, we use the ground-truth annotation to generate accurate input regions and optimize network weights. On the testing stage, we fix the network parameters and update the segmentation results in an iterative manner. We evaluate our approach on the NIH pancreas segmentation dataset, and outperform the state-of-the-art by more than 4{\%}, measured by the average Dice-S{\o}rensen Coefficient (DSC). In addition, we report 62.43{\%} DSC in the worst case, which guarantees the reliability of our approach in clinical applications.",
author = "Yuyin Zhou and Lingxi Xie and Wei Shen and Yan Wang and Fishman, {Elliot K} and Yuille, {Alan L.}",
year = "2017",
doi = "10.1007/978-3-319-66182-7_79",
language = "English (US)",
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 = "693--701",
booktitle = "Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings",

}

TY - GEN

T1 - A fixed-point model for pancreas segmentation in abdominal CT scans

AU - Zhou, Yuyin

AU - Xie, Lingxi

AU - Shen, Wei

AU - Wang, Yan

AU - Fishman, Elliot K

AU - Yuille, Alan L.

PY - 2017

Y1 - 2017

N2 - Deep neural networks have been widely adopted for automatic organ segmentation from abdominal CT scans. However, the segmentation accuracy of some small organs (e.g., the pancreas) is sometimes below satisfaction, arguably because deep networks are easily disrupted by the complex and variable background regions which occupies a large fraction of the input volume. In this paper, we formulate this problem into a fixed-point model which uses a predicted segmentation mask to shrink the input region. This is motivated by the fact that a smaller input region often leads to more accurate segmentation. In the training process, we use the ground-truth annotation to generate accurate input regions and optimize network weights. On the testing stage, we fix the network parameters and update the segmentation results in an iterative manner. We evaluate our approach on the NIH pancreas segmentation dataset, and outperform the state-of-the-art by more than 4%, measured by the average Dice-Sørensen Coefficient (DSC). In addition, we report 62.43% DSC in the worst case, which guarantees the reliability of our approach in clinical applications.

AB - Deep neural networks have been widely adopted for automatic organ segmentation from abdominal CT scans. However, the segmentation accuracy of some small organs (e.g., the pancreas) is sometimes below satisfaction, arguably because deep networks are easily disrupted by the complex and variable background regions which occupies a large fraction of the input volume. In this paper, we formulate this problem into a fixed-point model which uses a predicted segmentation mask to shrink the input region. This is motivated by the fact that a smaller input region often leads to more accurate segmentation. In the training process, we use the ground-truth annotation to generate accurate input regions and optimize network weights. On the testing stage, we fix the network parameters and update the segmentation results in an iterative manner. We evaluate our approach on the NIH pancreas segmentation dataset, and outperform the state-of-the-art by more than 4%, measured by the average Dice-Sørensen Coefficient (DSC). In addition, we report 62.43% DSC in the worst case, which guarantees the reliability of our approach in clinical applications.

UR - http://www.scopus.com/inward/record.url?scp=85029380997&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029380997&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-66182-7_79

DO - 10.1007/978-3-319-66182-7_79

M3 - Conference contribution

AN - SCOPUS:85029380997

SN - 9783319661810

VL - 10433 LNCS

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

SP - 693

EP - 701

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

PB - Springer Verlag

ER -