Standard

End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging. / Snaauw, Gerard; Gong, Dong; Maicas, Gabriel; van den Hengel, Anton; Niessen, Wiro; Verjans, Johan; Carneiro, Gustavo.

ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging. IEEE, 2018. p. 802-805 8759276.

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

Harvard

Snaauw, G, Gong, D, Maicas, G, van den Hengel, A, Niessen, W, Verjans, J & Carneiro, G 2018, End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging. in ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging., 8759276, IEEE, pp. 802-805, IEEE International Symposium on Biomedical Imaging, ISBI 2019, Venice, Italy, 8/04/19. https://doi.org/10.1109/ISBI.2019.8759276

APA

Snaauw, G., Gong, D., Maicas, G., van den Hengel, A., Niessen, W., Verjans, J., & Carneiro, G. (2018). End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging. In ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging (pp. 802-805). [8759276] IEEE. https://doi.org/10.1109/ISBI.2019.8759276

Vancouver

Snaauw G, Gong D, Maicas G, van den Hengel A, Niessen W, Verjans J et al. End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging. In ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging. IEEE. 2018. p. 802-805. 8759276 https://doi.org/10.1109/ISBI.2019.8759276

Author

Snaauw, Gerard ; Gong, Dong ; Maicas, Gabriel ; van den Hengel, Anton ; Niessen, Wiro ; Verjans, Johan ; Carneiro, Gustavo. / End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging. ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging. IEEE, 2018. pp. 802-805

BibTeX

@inproceedings{0ff796ad20244f81a3dbab760ad739f5,
title = "End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging",
abstract = "Cardiac magnetic resonance (CMR) is used extensively in the diagnosis and management of cardiovascular disease. Deep learning methods have proven to deliver segmentation results comparable tohuman experts in CMR imaging, but there have been no convincing results for the problem of end-to-end segmentation and diagnosis from CMR. This is in part due to a lack of sufficiently large datasets required to train robust diagnosis models. In this paper, we proposea learning method to train diagnosis models, where our approach isdesigned to work with relatively small datasets. In particular, the optimisation loss is based on multi-task learning that jointly trains for the tasks of segmentation and diagnosis classification. We hypothesize that segmentation has a regularizing effect on the learning of features relevant for diagnosis. Using the 100 training and 50 testingsamples available from the Automated Cardiac Diagnosis Challenge (ACDC) dataset, which has a balanced distribution of 5 cardiac diagnoses, we observe a reduction of the classification error from 32{\%} to 22{\%}, and a faster convergence compared to a baseline without segmentation. To the best of our knowledge, this is the best diagnosis results from CMR using an end-to-end diagnosis and segmentation learning method.",
author = "Gerard Snaauw and Dong Gong and Gabriel Maicas and {van den Hengel}, Anton and Wiro Niessen and Johan Verjans and Gustavo Carneiro",
year = "2018",
doi = "10.1109/ISBI.2019.8759276",
language = "English",
pages = "802--805",
booktitle = "ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging",
publisher = "IEEE",
address = "United States",

}

RIS

TY - GEN

T1 - End-to-end diagnosis and segmentation learning from cardiac magnetic resonance imaging

AU - Snaauw, Gerard

AU - Gong, Dong

AU - Maicas, Gabriel

AU - van den Hengel, Anton

AU - Niessen, Wiro

AU - Verjans, Johan

AU - Carneiro, Gustavo

PY - 2018

Y1 - 2018

N2 - Cardiac magnetic resonance (CMR) is used extensively in the diagnosis and management of cardiovascular disease. Deep learning methods have proven to deliver segmentation results comparable tohuman experts in CMR imaging, but there have been no convincing results for the problem of end-to-end segmentation and diagnosis from CMR. This is in part due to a lack of sufficiently large datasets required to train robust diagnosis models. In this paper, we proposea learning method to train diagnosis models, where our approach isdesigned to work with relatively small datasets. In particular, the optimisation loss is based on multi-task learning that jointly trains for the tasks of segmentation and diagnosis classification. We hypothesize that segmentation has a regularizing effect on the learning of features relevant for diagnosis. Using the 100 training and 50 testingsamples available from the Automated Cardiac Diagnosis Challenge (ACDC) dataset, which has a balanced distribution of 5 cardiac diagnoses, we observe a reduction of the classification error from 32% to 22%, and a faster convergence compared to a baseline without segmentation. To the best of our knowledge, this is the best diagnosis results from CMR using an end-to-end diagnosis and segmentation learning method.

AB - Cardiac magnetic resonance (CMR) is used extensively in the diagnosis and management of cardiovascular disease. Deep learning methods have proven to deliver segmentation results comparable tohuman experts in CMR imaging, but there have been no convincing results for the problem of end-to-end segmentation and diagnosis from CMR. This is in part due to a lack of sufficiently large datasets required to train robust diagnosis models. In this paper, we proposea learning method to train diagnosis models, where our approach isdesigned to work with relatively small datasets. In particular, the optimisation loss is based on multi-task learning that jointly trains for the tasks of segmentation and diagnosis classification. We hypothesize that segmentation has a regularizing effect on the learning of features relevant for diagnosis. Using the 100 training and 50 testingsamples available from the Automated Cardiac Diagnosis Challenge (ACDC) dataset, which has a balanced distribution of 5 cardiac diagnoses, we observe a reduction of the classification error from 32% to 22%, and a faster convergence compared to a baseline without segmentation. To the best of our knowledge, this is the best diagnosis results from CMR using an end-to-end diagnosis and segmentation learning method.

UR - http://www.scopus.com/inward/record.url?scp=85070410019&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2019.8759276

DO - 10.1109/ISBI.2019.8759276

M3 - Conference contribution

SP - 802

EP - 805

BT - ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging

PB - IEEE

ER -

ID: 47535739