King's College London

Research portal

On the compactness, efficiency, and representation of 3D convolutional networks: Brain parcellation as a pretext task

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

Standard

On the compactness, efficiency, and representation of 3D convolutional networks : Brain parcellation as a pretext task. / Li, Wenqi; Wang, Guotai; Fidon, Lucas; Ourselin, Sebastien; Cardoso, M. Jorge; Vercauteren, Tom.

Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. Vol. 10265 LNCS Springer Verlag, 2017. p. 348-360 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10265 LNCS).

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

Harvard

Li, W, Wang, G, Fidon, L, Ourselin, S, Cardoso, MJ & Vercauteren, T 2017, On the compactness, efficiency, and representation of 3D convolutional networks: Brain parcellation as a pretext task. in Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. vol. 10265 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10265 LNCS, Springer Verlag, pp. 348-360, 25th International Conference on Information Processing in Medical Imaging, IPMI 2017, Boone, United States, 25/06/2017. https://doi.org/10.1007/978-3-319-59050-9_28

APA

Li, W., Wang, G., Fidon, L., Ourselin, S., Cardoso, M. J., & Vercauteren, T. (2017). On the compactness, efficiency, and representation of 3D convolutional networks: Brain parcellation as a pretext task. In Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings (Vol. 10265 LNCS, pp. 348-360). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10265 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-59050-9_28

Vancouver

Li W, Wang G, Fidon L, Ourselin S, Cardoso MJ, Vercauteren T. On the compactness, efficiency, and representation of 3D convolutional networks: Brain parcellation as a pretext task. In Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. Vol. 10265 LNCS. Springer Verlag. 2017. p. 348-360. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-59050-9_28

Author

Li, Wenqi ; Wang, Guotai ; Fidon, Lucas ; Ourselin, Sebastien ; Cardoso, M. Jorge ; Vercauteren, Tom. / On the compactness, efficiency, and representation of 3D convolutional networks : Brain parcellation as a pretext task. Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. Vol. 10265 LNCS Springer Verlag, 2017. pp. 348-360 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

Bibtex Download

@inbook{340a9d5e76f54752bede0a7dd3cfe0f2,
title = "On the compactness, efficiency, and representation of 3D convolutional networks: Brain parcellation as a pretext task",
abstract = "Deep convolutional neural networks are powerful tools for learning visual representations from images. However, designing efficient deep architectures to analyse volumetric medical images remains challenging. This work investigates efficient and flexible elements of modern convolutional networks such as dilated convolution and residual connection. With these essential building blocks, we propose a high-resolution, compact convolutional network for volumetric image segmentation. To illustrate its efficiency of learning 3D representation from large-scale image data, the proposed network is validated with the challenging task of parcellating 155 neuroanatomical structures from brain MR images. Our experiments show that the proposed network architecture compares favourably with state-of-the-art volumetric segmentation networks while being an order of magnitude more compact. We consider the brain parcellation task as a pretext task for volumetric image segmentation; our trained network potentially provides a good starting point for transfer learning. Additionally, we show the feasibility of voxel-level uncertainty estimation using a sampling approximation through dropout.",
author = "Wenqi Li and Guotai Wang and Lucas Fidon and Sebastien Ourselin and Cardoso, {M. Jorge} and Tom Vercauteren",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-59050-9_28",
language = "English",
isbn = "9783319590493",
volume = "10265 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "348--360",
booktitle = "Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings",
address = "Germany",

}

RIS (suitable for import to EndNote) Download

TY - CHAP

T1 - On the compactness, efficiency, and representation of 3D convolutional networks

T2 - Brain parcellation as a pretext task

AU - Li, Wenqi

AU - Wang, Guotai

AU - Fidon, Lucas

AU - Ourselin, Sebastien

AU - Cardoso, M. Jorge

AU - Vercauteren, Tom

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Deep convolutional neural networks are powerful tools for learning visual representations from images. However, designing efficient deep architectures to analyse volumetric medical images remains challenging. This work investigates efficient and flexible elements of modern convolutional networks such as dilated convolution and residual connection. With these essential building blocks, we propose a high-resolution, compact convolutional network for volumetric image segmentation. To illustrate its efficiency of learning 3D representation from large-scale image data, the proposed network is validated with the challenging task of parcellating 155 neuroanatomical structures from brain MR images. Our experiments show that the proposed network architecture compares favourably with state-of-the-art volumetric segmentation networks while being an order of magnitude more compact. We consider the brain parcellation task as a pretext task for volumetric image segmentation; our trained network potentially provides a good starting point for transfer learning. Additionally, we show the feasibility of voxel-level uncertainty estimation using a sampling approximation through dropout.

AB - Deep convolutional neural networks are powerful tools for learning visual representations from images. However, designing efficient deep architectures to analyse volumetric medical images remains challenging. This work investigates efficient and flexible elements of modern convolutional networks such as dilated convolution and residual connection. With these essential building blocks, we propose a high-resolution, compact convolutional network for volumetric image segmentation. To illustrate its efficiency of learning 3D representation from large-scale image data, the proposed network is validated with the challenging task of parcellating 155 neuroanatomical structures from brain MR images. Our experiments show that the proposed network architecture compares favourably with state-of-the-art volumetric segmentation networks while being an order of magnitude more compact. We consider the brain parcellation task as a pretext task for volumetric image segmentation; our trained network potentially provides a good starting point for transfer learning. Additionally, we show the feasibility of voxel-level uncertainty estimation using a sampling approximation through dropout.

UR - http://www.scopus.com/inward/record.url?scp=85020488981&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-59050-9_28

DO - 10.1007/978-3-319-59050-9_28

M3 - Conference paper

AN - SCOPUS:85020488981

SN - 9783319590493

VL - 10265 LNCS

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

SP - 348

EP - 360

BT - Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings

PB - Springer Verlag

ER -

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454