Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Standard

Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning. / Xu, Peidi; Moshfeghifar, Faezeh; Gholamalizadeh, Torkan; Nielsen, Michael Bachmann; Erleben, Kenny; Darkner, Sune.

Medical Image Learning with Limited and Noisy Data: First International Workshop, MILLanD 2022 Held in Conjunction with MICCAI 2022 Singapore, September 22, 2022 Proceedings. ed. / Ghada Zamzmi; Sameer Antani; Sivaramakrishnan Rajaraman; Zhiyun Xue; Ulas Bagci; Marius George Linguraru. Springer Science and Business Media Deutschland GmbH, 2022. p. 153-162 (Medical Image Learning with Limited and Noisy Data, Vol. 13559).

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Harvard

Xu, P, Moshfeghifar, F, Gholamalizadeh, T, Nielsen, MB, Erleben, K & Darkner, S 2022, Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning. in G Zamzmi, S Antani, S Rajaraman, Z Xue, U Bagci & MG Linguraru (eds), Medical Image Learning with Limited and Noisy Data: First International Workshop, MILLanD 2022 Held in Conjunction with MICCAI 2022 Singapore, September 22, 2022 Proceedings. Springer Science and Business Media Deutschland GmbH, Medical Image Learning with Limited and Noisy Data, vol. 13559, pp. 153-162, 1st International Workshop on Medical Image Learning with Limited and Noisy Data, MILLanD 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2022, Singapore, Singapore, 22/09/2022. https://doi.org/10.1007/978-3-031-16760-7_15

APA

Xu, P., Moshfeghifar, F., Gholamalizadeh, T., Nielsen, M. B., Erleben, K., & Darkner, S. (2022). Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning. In G. Zamzmi, S. Antani, S. Rajaraman, Z. Xue, U. Bagci, & M. G. Linguraru (Eds.), Medical Image Learning with Limited and Noisy Data: First International Workshop, MILLanD 2022 Held in Conjunction with MICCAI 2022 Singapore, September 22, 2022 Proceedings (pp. 153-162). Springer Science and Business Media Deutschland GmbH. Medical Image Learning with Limited and Noisy Data Vol. 13559 https://doi.org/10.1007/978-3-031-16760-7_15

Vancouver

Xu P, Moshfeghifar F, Gholamalizadeh T, Nielsen MB, Erleben K, Darkner S. Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning. In Zamzmi G, Antani S, Rajaraman S, Xue Z, Bagci U, Linguraru MG, editors, Medical Image Learning with Limited and Noisy Data: First International Workshop, MILLanD 2022 Held in Conjunction with MICCAI 2022 Singapore, September 22, 2022 Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 153-162. (Medical Image Learning with Limited and Noisy Data, Vol. 13559). https://doi.org/10.1007/978-3-031-16760-7_15

Author

Xu, Peidi ; Moshfeghifar, Faezeh ; Gholamalizadeh, Torkan ; Nielsen, Michael Bachmann ; Erleben, Kenny ; Darkner, Sune. / Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning. Medical Image Learning with Limited and Noisy Data: First International Workshop, MILLanD 2022 Held in Conjunction with MICCAI 2022 Singapore, September 22, 2022 Proceedings. editor / Ghada Zamzmi ; Sameer Antani ; Sivaramakrishnan Rajaraman ; Zhiyun Xue ; Ulas Bagci ; Marius George Linguraru. Springer Science and Business Media Deutschland GmbH, 2022. pp. 153-162 (Medical Image Learning with Limited and Noisy Data, Vol. 13559).

Bibtex

@inproceedings{d5a3e7533871484ba673f7aae4673525,
title = "Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning",
abstract = "Accurate geometry representation is essential in developing finite element models. Although generally good, deep-learning segmentation approaches with only few data have difficulties in accurately segmenting fine features, e.g., gaps and thin structures. Subsequently, segmented geometries need labor-intensive manual modifications to reach a quality where they can be used for simulation purposes. We propose a strategy that uses transfer learning to reuse datasets with poor segmentation combined with an interactive learning step where fine-tuning of the data results in anatomically accurate segmentations suitable for simulations. We use a modified MultiPlanar UNet that is pre-trained using inferior hip joint segmentation combined with a dedicated loss function to learn the gap regions and post-processing to correct tiny inaccuracies on symmetric classes due to rotational invariance. We demonstrate this robust yet conceptually simple approach applied with clinically validated results on publicly available computed tomography scans of hip joints. Code and resulting 3D models are available at: https://github.com/MICCAI2022-155/AuToSeg.",
keywords = "Finite element modeling, Segmentation, Transfer learning",
author = "Peidi Xu and Faezeh Moshfeghifar and Torkan Gholamalizadeh and Nielsen, {Michael Bachmann} and Kenny Erleben and Sune Darkner",
note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 1st International Workshop on Medical Image Learning with Limited and Noisy Data, MILLanD 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2022 ; Conference date: 22-09-2022 Through 22-09-2022",
year = "2022",
doi = "10.1007/978-3-031-16760-7_15",
language = "English",
isbn = " 978-3-031-16759-1",
series = "Medical Image Learning with Limited and Noisy Data",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "153--162",
editor = "Ghada Zamzmi and Sameer Antani and Sivaramakrishnan Rajaraman and Zhiyun Xue and Ulas Bagci and Linguraru, {Marius George}",
booktitle = "Medical Image Learning with Limited and Noisy Data",
address = "Germany",

}

RIS

TY - GEN

T1 - Auto-segmentation of Hip Joints Using MultiPlanar UNet with Transfer Learning

AU - Xu, Peidi

AU - Moshfeghifar, Faezeh

AU - Gholamalizadeh, Torkan

AU - Nielsen, Michael Bachmann

AU - Erleben, Kenny

AU - Darkner, Sune

N1 - Publisher Copyright: © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

PY - 2022

Y1 - 2022

N2 - Accurate geometry representation is essential in developing finite element models. Although generally good, deep-learning segmentation approaches with only few data have difficulties in accurately segmenting fine features, e.g., gaps and thin structures. Subsequently, segmented geometries need labor-intensive manual modifications to reach a quality where they can be used for simulation purposes. We propose a strategy that uses transfer learning to reuse datasets with poor segmentation combined with an interactive learning step where fine-tuning of the data results in anatomically accurate segmentations suitable for simulations. We use a modified MultiPlanar UNet that is pre-trained using inferior hip joint segmentation combined with a dedicated loss function to learn the gap regions and post-processing to correct tiny inaccuracies on symmetric classes due to rotational invariance. We demonstrate this robust yet conceptually simple approach applied with clinically validated results on publicly available computed tomography scans of hip joints. Code and resulting 3D models are available at: https://github.com/MICCAI2022-155/AuToSeg.

AB - Accurate geometry representation is essential in developing finite element models. Although generally good, deep-learning segmentation approaches with only few data have difficulties in accurately segmenting fine features, e.g., gaps and thin structures. Subsequently, segmented geometries need labor-intensive manual modifications to reach a quality where they can be used for simulation purposes. We propose a strategy that uses transfer learning to reuse datasets with poor segmentation combined with an interactive learning step where fine-tuning of the data results in anatomically accurate segmentations suitable for simulations. We use a modified MultiPlanar UNet that is pre-trained using inferior hip joint segmentation combined with a dedicated loss function to learn the gap regions and post-processing to correct tiny inaccuracies on symmetric classes due to rotational invariance. We demonstrate this robust yet conceptually simple approach applied with clinically validated results on publicly available computed tomography scans of hip joints. Code and resulting 3D models are available at: https://github.com/MICCAI2022-155/AuToSeg.

KW - Finite element modeling

KW - Segmentation

KW - Transfer learning

U2 - 10.1007/978-3-031-16760-7_15

DO - 10.1007/978-3-031-16760-7_15

M3 - Article in proceedings

SN - 978-3-031-16759-1

T3 - Medical Image Learning with Limited and Noisy Data

SP - 153

EP - 162

BT - Medical Image Learning with Limited and Noisy Data

A2 - Zamzmi, Ghada

A2 - Antani, Sameer

A2 - Rajaraman, Sivaramakrishnan

A2 - Xue, Zhiyun

A2 - Bagci, Ulas

A2 - Linguraru, Marius George

PB - Springer Science and Business Media Deutschland GmbH

T2 - 1st International Workshop on Medical Image Learning with Limited and Noisy Data, MILLanD 2022, held in conjunction with 25th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2022

Y2 - 22 September 2022 through 22 September 2022

ER -

ID: 320498297