File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Class-balanced deep neural network for automatic ventricular structure segmentation

TitleClass-balanced deep neural network for automatic ventricular structure segmentation
Authors
Issue Date2018
PublisherSpringer.
Citation
8th International Workshop on Statistical Atlases and Computational Models of the Heart (STACOM 2017), Held in Conjunction with MICCAI 2017, Quebec City, Canada, 10-14 September 2017. In Pop, M, Sermesant, M, Jodoin, P, et al. (Eds.), Statistical Atlases and Computational Models of the Heart. ACDC and MMWHS Challenges: 8th International Workshop, STACOM 2017, Held in Conjunction with MICCAI 2017, Quebec City, Canada, September 10-14, 2017, Revised Selected Papers, p. 152-160. Cham, Switzerland: Springer, 2018 How to Cite?
AbstractSegmenting ventricular structures from cardiovascular MR scan is important for quantitative evaluation of heart. Manual delineation is time-consuming and tedious and lack of reproductivity. Considering MR image quality, heart variance, spatial inconsistency and motion artifacts during scanning, it is still a non-trivial task for automatic segmentation methods. In this paper, we propose a general and fully automatic solution to concurrently segment three important ventricular structures. Rooting in the deep learning trend, our method starts from 3D Fully Convolutional Network (3D FCN). We then enhance the 3D FCN with two well-verified blocks: (1) we conduct transfer learning between a pre-trained C3D model and our 3D FCN to get good initialization and thus suppress overfitting. (2) since boosting the gradient flow in network is beneficial to promote segmentation performance, we attach several auxiliary loss functions so as to expose early layers to better supervision. Because the volume size imbalance among different ventricular structures often biases the training of our 3D FCN, to this end, we investigate the capacity of different loss functions and propose a Multi-class Dice Similarity Coefficient (mDSC) based loss function to re-weight the training for all classes. We verified our method, especially the significance of mDSC, on the Automated Cardiac Diagnosis Challenge 2017 datasets for MR image segmentation. Extensive experimental results demonstrate the promising performance of our method.
Persistent Identifierhttp://hdl.handle.net/10722/299569
ISBN
ISSN
2023 SCImago Journal Rankings: 0.606
ISI Accession Number ID
Series/Report no.Lecture Notes in Computer Science ; 10663

 

DC FieldValueLanguage
dc.contributor.authorYang, Xin-
dc.contributor.authorBian, Cheng-
dc.contributor.authorYu, Lequan-
dc.contributor.authorNi, Dong-
dc.contributor.authorHeng, Pheng Ann-
dc.date.accessioned2021-05-21T03:34:41Z-
dc.date.available2021-05-21T03:34:41Z-
dc.date.issued2018-
dc.identifier.citation8th International Workshop on Statistical Atlases and Computational Models of the Heart (STACOM 2017), Held in Conjunction with MICCAI 2017, Quebec City, Canada, 10-14 September 2017. In Pop, M, Sermesant, M, Jodoin, P, et al. (Eds.), Statistical Atlases and Computational Models of the Heart. ACDC and MMWHS Challenges: 8th International Workshop, STACOM 2017, Held in Conjunction with MICCAI 2017, Quebec City, Canada, September 10-14, 2017, Revised Selected Papers, p. 152-160. Cham, Switzerland: Springer, 2018-
dc.identifier.isbn9783319755403-
dc.identifier.issn0302-9743-
dc.identifier.urihttp://hdl.handle.net/10722/299569-
dc.description.abstractSegmenting ventricular structures from cardiovascular MR scan is important for quantitative evaluation of heart. Manual delineation is time-consuming and tedious and lack of reproductivity. Considering MR image quality, heart variance, spatial inconsistency and motion artifacts during scanning, it is still a non-trivial task for automatic segmentation methods. In this paper, we propose a general and fully automatic solution to concurrently segment three important ventricular structures. Rooting in the deep learning trend, our method starts from 3D Fully Convolutional Network (3D FCN). We then enhance the 3D FCN with two well-verified blocks: (1) we conduct transfer learning between a pre-trained C3D model and our 3D FCN to get good initialization and thus suppress overfitting. (2) since boosting the gradient flow in network is beneficial to promote segmentation performance, we attach several auxiliary loss functions so as to expose early layers to better supervision. Because the volume size imbalance among different ventricular structures often biases the training of our 3D FCN, to this end, we investigate the capacity of different loss functions and propose a Multi-class Dice Similarity Coefficient (mDSC) based loss function to re-weight the training for all classes. We verified our method, especially the significance of mDSC, on the Automated Cardiac Diagnosis Challenge 2017 datasets for MR image segmentation. Extensive experimental results demonstrate the promising performance of our method.-
dc.languageeng-
dc.publisherSpringer.-
dc.relation.ispartofStatistical Atlases and Computational Models of the Heart. ACDC and MMWHS Challenges: 8th International Workshop, STACOM 2017, Held in Conjunction with MICCAI 2017, Quebec City, Canada, September 10-14, 2017, Revised Selected Papers-
dc.relation.ispartofseriesLecture Notes in Computer Science ; 10663-
dc.titleClass-balanced deep neural network for automatic ventricular structure segmentation-
dc.typeConference_Paper-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1007/978-3-319-75541-0_16-
dc.identifier.scopuseid_2-s2.0-85044437048-
dc.identifier.spage152-
dc.identifier.epage160-
dc.identifier.eissn1611-3349-
dc.identifier.isiWOS:000550266300016-
dc.publisher.placeCham, Switzerland-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats