Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

5 Citations (Scopus)

Abstract

Successful automated detection of short needles during an intervention is necessary to allow the physician identify and correct any misalignment of the needle and the target at early stages, which reduces needle passes and improves health outcomes. In this paper, we present a novel approach to detect needle voxels in 3D ultrasound volume with high precision using convolutional neural networks. Each voxel is classified from locally-extracted raw data of three orthogonal planes centered on it. We propose a bootstrap re-sampling approach to enhance the training in our highly imbalanced data. The proposed method successfully detects 17G and 22G needles with a single trained network, showing a robust generalized approach. Extensive ex-vivo evaluations on 3D ultrasound datasets of chicken breast show 25% increase in F1-score over the state-of-the-art feature-based method. Furthermore, very short needles inserted for only 5 mm in the volume are detected with tip localization errors of <0.5 mm, indicating that the tip is always visible in the detected plane.
LanguageEnglish
Title of host publicationMedical Image Computing and Computer-Assisted Intervention − MICCAI 2017
Subtitle of host publication20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part II
Place of PublicationDordrecht
PublisherSpringer
Pages610-618
ISBN (Electronic)978-3-319-66185-8
ISBN (Print)978-3-319-66184-1
DOIs
StatePublished - 4 Sep 2017

Publication series

NameLNCS
Volume10434

Fingerprint

Needles
Ultrasonics
Health
Sampling
Neural networks

Cite this

Pourtaherian, A., Ghazvinian Zanjani, F., Zinger, S., Mihajlovic, N., Ng, G. C., Korsten, H. H. M., & de With, P. H. N. (2017). Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks. In Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part II (pp. 610-618). (LNCS; Vol. 10434). Dordrecht: Springer. DOI: 10.1007/978-3-319-66185-8_69
Pourtaherian, A. ; Ghazvinian Zanjani, F. ; Zinger, S. ; Mihajlovic, N. ; Ng, G.C. ; Korsten, H.H.M. ; de With, P.H.N./ Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks. Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part II. Dordrecht : Springer, 2017. pp. 610-618 (LNCS).
@inproceedings{d4192090d7e74a17b07dd48f29816fba,
title = "Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks",
abstract = "Successful automated detection of short needles during an intervention is necessary to allow the physician identify and correct any misalignment of the needle and the target at early stages, which reduces needle passes and improves health outcomes. In this paper, we present a novel approach to detect needle voxels in 3D ultrasound volume with high precision using convolutional neural networks. Each voxel is classified from locally-extracted raw data of three orthogonal planes centered on it. We propose a bootstrap re-sampling approach to enhance the training in our highly imbalanced data. The proposed method successfully detects 17G and 22G needles with a single trained network, showing a robust generalized approach. Extensive ex-vivo evaluations on 3D ultrasound datasets of chicken breast show 25{\%} increase in F1-score over the state-of-the-art feature-based method. Furthermore, very short needles inserted for only 5 mm in the volume are detected with tip localization errors of <0.5 mm, indicating that the tip is always visible in the detected plane.",
author = "A. Pourtaherian and {Ghazvinian Zanjani}, F. and S. Zinger and N. Mihajlovic and G.C. Ng and H.H.M. Korsten and {de With}, P.H.N.",
year = "2017",
month = "9",
day = "4",
doi = "10.1007/978-3-319-66185-8_69",
language = "English",
isbn = "978-3-319-66184-1",
series = "LNCS",
publisher = "Springer",
pages = "610--618",
booktitle = "Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017",
address = "Germany",

}

Pourtaherian, A, Ghazvinian Zanjani, F, Zinger, S, Mihajlovic, N, Ng, GC, Korsten, HHM & de With, PHN 2017, Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks. in Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part II. LNCS, vol. 10434, Springer, Dordrecht, pp. 610-618. DOI: 10.1007/978-3-319-66185-8_69

Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks. / Pourtaherian, A.; Ghazvinian Zanjani, F.; Zinger, S.; Mihajlovic, N.; Ng, G.C.; Korsten, H.H.M.; de With, P.H.N.

Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part II. Dordrecht : Springer, 2017. p. 610-618 (LNCS; Vol. 10434).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks

AU - Pourtaherian,A.

AU - Ghazvinian Zanjani,F.

AU - Zinger,S.

AU - Mihajlovic,N.

AU - Ng,G.C.

AU - Korsten,H.H.M.

AU - de With,P.H.N.

PY - 2017/9/4

Y1 - 2017/9/4

N2 - Successful automated detection of short needles during an intervention is necessary to allow the physician identify and correct any misalignment of the needle and the target at early stages, which reduces needle passes and improves health outcomes. In this paper, we present a novel approach to detect needle voxels in 3D ultrasound volume with high precision using convolutional neural networks. Each voxel is classified from locally-extracted raw data of three orthogonal planes centered on it. We propose a bootstrap re-sampling approach to enhance the training in our highly imbalanced data. The proposed method successfully detects 17G and 22G needles with a single trained network, showing a robust generalized approach. Extensive ex-vivo evaluations on 3D ultrasound datasets of chicken breast show 25% increase in F1-score over the state-of-the-art feature-based method. Furthermore, very short needles inserted for only 5 mm in the volume are detected with tip localization errors of <0.5 mm, indicating that the tip is always visible in the detected plane.

AB - Successful automated detection of short needles during an intervention is necessary to allow the physician identify and correct any misalignment of the needle and the target at early stages, which reduces needle passes and improves health outcomes. In this paper, we present a novel approach to detect needle voxels in 3D ultrasound volume with high precision using convolutional neural networks. Each voxel is classified from locally-extracted raw data of three orthogonal planes centered on it. We propose a bootstrap re-sampling approach to enhance the training in our highly imbalanced data. The proposed method successfully detects 17G and 22G needles with a single trained network, showing a robust generalized approach. Extensive ex-vivo evaluations on 3D ultrasound datasets of chicken breast show 25% increase in F1-score over the state-of-the-art feature-based method. Furthermore, very short needles inserted for only 5 mm in the volume are detected with tip localization errors of <0.5 mm, indicating that the tip is always visible in the detected plane.

U2 - 10.1007/978-3-319-66185-8_69

DO - 10.1007/978-3-319-66185-8_69

M3 - Conference contribution

SN - 978-3-319-66184-1

T3 - LNCS

SP - 610

EP - 618

BT - Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017

PB - Springer

CY - Dordrecht

ER -

Pourtaherian A, Ghazvinian Zanjani F, Zinger S, Mihajlovic N, Ng GC, Korsten HHM et al. Improving needle detection in 3D ultrasound using orthogonal-plane convolutional networks. In Medical Image Computing and Computer-Assisted Intervention − MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings, Part II. Dordrecht: Springer. 2017. p. 610-618. (LNCS). Available from, DOI: 10.1007/978-3-319-66185-8_69