Aarhus Universitets segl

Predicting embryo viability based on self-supervised alignment of time-lapse videos

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

Standard

Predicting embryo viability based on self-supervised alignment of time-lapse videos. / Kragh, Mikkel Fly; Rimestad, Jens; Lassen, Jacob Theilgaard et al.

I: IEEE Transactions on Medical Imaging, Bind 41, Nr. 2, 02.2022, s. 465-475.

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

Harvard

Kragh, MF, Rimestad, J, Lassen, JT, Berntsen, J & Karstoft, H 2022, 'Predicting embryo viability based on self-supervised alignment of time-lapse videos', IEEE Transactions on Medical Imaging, bind 41, nr. 2, s. 465-475. https://doi.org/10.1109/TMI.2021.3116986

APA

Kragh, M. F., Rimestad, J., Lassen, J. T., Berntsen, J., & Karstoft, H. (2022). Predicting embryo viability based on self-supervised alignment of time-lapse videos. IEEE Transactions on Medical Imaging, 41(2), 465-475. https://doi.org/10.1109/TMI.2021.3116986

CBE

Kragh MF, Rimestad J, Lassen JT, Berntsen J, Karstoft H. 2022. Predicting embryo viability based on self-supervised alignment of time-lapse videos. IEEE Transactions on Medical Imaging. 41(2):465-475. https://doi.org/10.1109/TMI.2021.3116986

MLA

Kragh, Mikkel Fly et al. "Predicting embryo viability based on self-supervised alignment of time-lapse videos". IEEE Transactions on Medical Imaging. 2022, 41(2). 465-475. https://doi.org/10.1109/TMI.2021.3116986

Vancouver

Kragh MF, Rimestad J, Lassen JT, Berntsen J, Karstoft H. Predicting embryo viability based on self-supervised alignment of time-lapse videos. IEEE Transactions on Medical Imaging. 2022 feb.;41(2):465-475. Epub 2021 okt. 1. doi: 10.1109/TMI.2021.3116986

Author

Kragh, Mikkel Fly ; Rimestad, Jens ; Lassen, Jacob Theilgaard et al. / Predicting embryo viability based on self-supervised alignment of time-lapse videos. I: IEEE Transactions on Medical Imaging. 2022 ; Bind 41, Nr. 2. s. 465-475.

Bibtex

@article{1748abfb63d949b6a4e3e87e565845f8,
title = "Predicting embryo viability based on self-supervised alignment of time-lapse videos",
abstract = "With self-supervised learning, both labeled and unlabeled data can be used for representation learning and model pretraining. This is particularly relevant when automating the selection of a patient's fertilized eggs (embryos) during a fertility treatment, in which only the embryos that were transferred to the female uterus may have labels of pregnancy. In this paper, we apply a self-supervised video alignment method known as temporal cycle-consistency (TCC) on 38176 time-lapse videos of developing embryos, of which 14550 were labeled. We show how TCC can be used to extract temporal similarities between embryo videos and use these for predicting pregnancy likelihood. Our temporal similarity method outperforms the time alignment measurement (TAM) with an area under the receiver operating characteristic (AUC) of 0.64 vs. 0.56. Compared to existing embryo evaluation models, it places in between a pure temporal and a spatio-temporal model that both require manual annotations. Furthermore, we use TCC for transfer learning in a semi-supervised fashion and show significant performance improvements compared to standard supervised learning, when only a small subset of the dataset is labeled. Specifically, two variants of transfer learning both achieve an AUC of 0.66 compared to 0.63 for supervised learning when 16% of the dataset is labeled.",
keywords = "Annotations, Clustering, Embryo, Embryo Selection, In vitro fertilization, Manuals, Pregnancy, Self-supervised learning, Supervised learning, Task analysis, Temporal Cycle-Consistency, Videos, Temporal cycle-consistency, Embryo selection, self-supervised learning, temporal cycle-consistency, in vitro fertilization, embryo selection",
author = "Kragh, {Mikkel Fly} and Jens Rimestad and Lassen, {Jacob Theilgaard} and J{\o}rgen Berntsen and Henrik Karstoft",
year = "2022",
month = feb,
doi = "10.1109/TMI.2021.3116986",
language = "English",
volume = "41",
pages = "465--475",
journal = "I E E E Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers",
number = "2",

}

RIS

TY - JOUR

T1 - Predicting embryo viability based on self-supervised alignment of time-lapse videos

AU - Kragh, Mikkel Fly

AU - Rimestad, Jens

AU - Lassen, Jacob Theilgaard

AU - Berntsen, Jørgen

AU - Karstoft, Henrik

PY - 2022/2

Y1 - 2022/2

N2 - With self-supervised learning, both labeled and unlabeled data can be used for representation learning and model pretraining. This is particularly relevant when automating the selection of a patient's fertilized eggs (embryos) during a fertility treatment, in which only the embryos that were transferred to the female uterus may have labels of pregnancy. In this paper, we apply a self-supervised video alignment method known as temporal cycle-consistency (TCC) on 38176 time-lapse videos of developing embryos, of which 14550 were labeled. We show how TCC can be used to extract temporal similarities between embryo videos and use these for predicting pregnancy likelihood. Our temporal similarity method outperforms the time alignment measurement (TAM) with an area under the receiver operating characteristic (AUC) of 0.64 vs. 0.56. Compared to existing embryo evaluation models, it places in between a pure temporal and a spatio-temporal model that both require manual annotations. Furthermore, we use TCC for transfer learning in a semi-supervised fashion and show significant performance improvements compared to standard supervised learning, when only a small subset of the dataset is labeled. Specifically, two variants of transfer learning both achieve an AUC of 0.66 compared to 0.63 for supervised learning when 16% of the dataset is labeled.

AB - With self-supervised learning, both labeled and unlabeled data can be used for representation learning and model pretraining. This is particularly relevant when automating the selection of a patient's fertilized eggs (embryos) during a fertility treatment, in which only the embryos that were transferred to the female uterus may have labels of pregnancy. In this paper, we apply a self-supervised video alignment method known as temporal cycle-consistency (TCC) on 38176 time-lapse videos of developing embryos, of which 14550 were labeled. We show how TCC can be used to extract temporal similarities between embryo videos and use these for predicting pregnancy likelihood. Our temporal similarity method outperforms the time alignment measurement (TAM) with an area under the receiver operating characteristic (AUC) of 0.64 vs. 0.56. Compared to existing embryo evaluation models, it places in between a pure temporal and a spatio-temporal model that both require manual annotations. Furthermore, we use TCC for transfer learning in a semi-supervised fashion and show significant performance improvements compared to standard supervised learning, when only a small subset of the dataset is labeled. Specifically, two variants of transfer learning both achieve an AUC of 0.66 compared to 0.63 for supervised learning when 16% of the dataset is labeled.

KW - Annotations

KW - Clustering

KW - Embryo

KW - Embryo Selection

KW - In vitro fertilization

KW - Manuals

KW - Pregnancy

KW - Self-supervised learning

KW - Supervised learning

KW - Task analysis

KW - Temporal Cycle-Consistency

KW - Videos

KW - Temporal cycle-consistency

KW - Embryo selection

KW - self-supervised learning

KW - temporal cycle-consistency

KW - in vitro fertilization

KW - embryo selection

U2 - 10.1109/TMI.2021.3116986

DO - 10.1109/TMI.2021.3116986

M3 - Journal article

C2 - 34596537

VL - 41

SP - 465

EP - 475

JO - I E E E Transactions on Medical Imaging

JF - I E E E Transactions on Medical Imaging

SN - 0278-0062

IS - 2

ER -