3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor

Helena Williams; Laura Cattani; Wenqi Li; Mahdi Tabassian; Tom Vercauteren; Jan Deprest; Jan D'Hooge

doi:10.1109/ULTSYM.2019.8925792

3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor

Helena Williams, Laura Cattani, Wenqi Li, Mahdi Tabassian, Tom Vercauteren, Jan Deprest, Jan D'Hooge

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

5 Citations (Scopus)

Abstract

Pelvic organ prolapse (POP) decreases the quality of life for many women. To assess POP, the levator hiatus is segmented in a 2D plane of minimal hiatal dimensions, known as the C-plane. In order to automate plane detection, landmark information of key structures should be given to a plane detection algorithm. In this work, we present a fully automatic method to segment the urethra from a 3D transperineal ultrasound volume using a convolutional neural network (CNN). A dataset with 35 volumes from 20 patients during the Valsalva manoeuver (i.e. Valsalva, contraction and rest) labelled by an expert, was used for training and evaluation in a 5-fold cross-validation process. The 3D CNN model yielded an average robust Hausdorff distance of 4.68mm (95 percentile) which was comparable to intra-observer results.

Original language	English
Title of host publication	2019 IEEE International Ultrasonics Symposium, IUS 2019
Publisher	IEEE Computer Society
Pages	1473-1476
Number of pages	4
Volume	2019-October
ISBN (Electronic)	9781728145969
DOIs	https://doi.org/10.1109/ULTSYM.2019.8925792
Publication status	Published - 1 Oct 2019
Event	2019 IEEE International Ultrasonics Symposium, IUS 2019 - Glasgow, United Kingdom Duration: 6 Oct 2019 → 9 Oct 2019

Conference

Conference	2019 IEEE International Ultrasonics Symposium, IUS 2019
Country/Territory	United Kingdom
City	Glasgow
Period	6/10/2019 → 9/10/2019

Keywords

3D Convolutional neural network
Pelvic floor
Semantic segmentation

Access to Document

10.1109/ULTSYM.2019.8925792

Cite this

Williams, H., Cattani, L., Li, W., Tabassian, M., Vercauteren, T., Deprest, J., & D'Hooge, J. (2019). 3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor. In 2019 IEEE International Ultrasonics Symposium, IUS 2019 (Vol. 2019-October, pp. 1473-1476). Article 8925792 IEEE Computer Society. https://doi.org/10.1109/ULTSYM.2019.8925792

@inbook{483f1a8199d94f35b3114da89ad652c5,

title = "3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor",

abstract = "Pelvic organ prolapse (POP) decreases the quality of life for many women. To assess POP, the levator hiatus is segmented in a 2D plane of minimal hiatal dimensions, known as the C-plane. In order to automate plane detection, landmark information of key structures should be given to a plane detection algorithm. In this work, we present a fully automatic method to segment the urethra from a 3D transperineal ultrasound volume using a convolutional neural network (CNN). A dataset with 35 volumes from 20 patients during the Valsalva manoeuver (i.e. Valsalva, contraction and rest) labelled by an expert, was used for training and evaluation in a 5-fold cross-validation process. The 3D CNN model yielded an average robust Hausdorff distance of 4.68mm (95 percentile) which was comparable to intra-observer results.",

keywords = "3D Convolutional neural network, Pelvic floor, Semantic segmentation",

author = "Helena Williams and Laura Cattani and Wenqi Li and Mahdi Tabassian and Tom Vercauteren and Jan Deprest and Jan D'Hooge",

year = "2019",

month = oct,

day = "1",

doi = "10.1109/ULTSYM.2019.8925792",

language = "English",

volume = "2019-October",

pages = "1473--1476",

booktitle = "2019 IEEE International Ultrasonics Symposium, IUS 2019",

publisher = "IEEE Computer Society",

note = "2019 IEEE International Ultrasonics Symposium, IUS 2019 ; Conference date: 06-10-2019 Through 09-10-2019",

}

Williams, H, Cattani, L, Li, W, Tabassian, M, Vercauteren, T, Deprest, J & D'Hooge, J 2019, 3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor. in 2019 IEEE International Ultrasonics Symposium, IUS 2019. vol. 2019-October, 8925792, IEEE Computer Society, pp. 1473-1476, 2019 IEEE International Ultrasonics Symposium, IUS 2019, Glasgow, United Kingdom, 6/10/2019. https://doi.org/10.1109/ULTSYM.2019.8925792

3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor. / Williams, Helena; Cattani, Laura; Li, Wenqi et al.
2019 IEEE International Ultrasonics Symposium, IUS 2019. Vol. 2019-October IEEE Computer Society, 2019. p. 1473-1476 8925792.

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

TY - CHAP

T1 - 3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor

AU - Williams, Helena

AU - Cattani, Laura

AU - Li, Wenqi

AU - Tabassian, Mahdi

AU - Vercauteren, Tom

AU - Deprest, Jan

AU - D'Hooge, Jan

PY - 2019/10/1

Y1 - 2019/10/1

N2 - Pelvic organ prolapse (POP) decreases the quality of life for many women. To assess POP, the levator hiatus is segmented in a 2D plane of minimal hiatal dimensions, known as the C-plane. In order to automate plane detection, landmark information of key structures should be given to a plane detection algorithm. In this work, we present a fully automatic method to segment the urethra from a 3D transperineal ultrasound volume using a convolutional neural network (CNN). A dataset with 35 volumes from 20 patients during the Valsalva manoeuver (i.e. Valsalva, contraction and rest) labelled by an expert, was used for training and evaluation in a 5-fold cross-validation process. The 3D CNN model yielded an average robust Hausdorff distance of 4.68mm (95 percentile) which was comparable to intra-observer results.

AB - Pelvic organ prolapse (POP) decreases the quality of life for many women. To assess POP, the levator hiatus is segmented in a 2D plane of minimal hiatal dimensions, known as the C-plane. In order to automate plane detection, landmark information of key structures should be given to a plane detection algorithm. In this work, we present a fully automatic method to segment the urethra from a 3D transperineal ultrasound volume using a convolutional neural network (CNN). A dataset with 35 volumes from 20 patients during the Valsalva manoeuver (i.e. Valsalva, contraction and rest) labelled by an expert, was used for training and evaluation in a 5-fold cross-validation process. The 3D CNN model yielded an average robust Hausdorff distance of 4.68mm (95 percentile) which was comparable to intra-observer results.

KW - 3D Convolutional neural network

KW - Pelvic floor

KW - Semantic segmentation

UR - http://www.scopus.com/inward/record.url?scp=85077568454&partnerID=8YFLogxK

U2 - 10.1109/ULTSYM.2019.8925792

DO - 10.1109/ULTSYM.2019.8925792

M3 - Conference paper

AN - SCOPUS:85077568454

VL - 2019-October

SP - 1473

EP - 1476

BT - 2019 IEEE International Ultrasonics Symposium, IUS 2019

PB - IEEE Computer Society

T2 - 2019 IEEE International Ultrasonics Symposium, IUS 2019

Y2 - 6 October 2019 through 9 October 2019

ER -

3D Convolutional Neural Network for Segmentation of the Urethra in Volumetric Ultrasound of the Pelvic Floor

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this