Training a neural network based on unreliable human annotation of medical images

Yair Dgani; Hayit Greenspan; Jacob Goldberger

doi:10.1109/ISBI.2018.8363518

Training a neural network based on unreliable human annotation of medical images

Yair Dgani, Hayit Greenspan, Jacob Goldberger

Tel Aviv University

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

59 Scopus citations

Abstract

Building classification models from clinical data often requires human experts for example labeling. However, it is difficult to obtain a perfect set of labels due to the complexity of the medical data and the large variability between experts. In this study we present a neural-network training strategy that is more robust to unreliable labeling by explicitly modeling the label noise as part of the network architecture. Our method is demonstrated on breast microcalcifications classification into benign and malignant categories, given multi-view mammograms. We show that the proposed training procedure outperforms standard training methods that ignore the existence of label noise.

Original language	English
Title of host publication	2018 IEEE 15th International Symposium on Biomedical Imaging, ISBI 2018
Publisher	IEEE Computer Society
Pages	39-42
Number of pages	4
ISBN (Electronic)	9781538636367
DOIs	https://doi.org/10.1109/ISBI.2018.8363518
State	Published - 23 May 2018
Event	15th IEEE International Symposium on Biomedical Imaging, ISBI 2018 - Washington, United States Duration: 4 Apr 2018 → 7 Apr 2018

Publication series

Name	Proceedings - International Symposium on Biomedical Imaging
Volume	2018-April
ISSN (Print)	1945-7928
ISSN (Electronic)	1945-8452

Conference

Conference	15th IEEE International Symposium on Biomedical Imaging, ISBI 2018
Country/Territory	United States
City	Washington
Period	4/04/18 → 7/04/18

Bibliographical note

Publisher Copyright:
© 2018 IEEE.

Keywords

Deep-learning
Mammography
Microcalcifications
Noisy-labels
Robust training

Access to Document

10.1109/ISBI.2018.8363518

Cite this

@inproceedings{8fff414271394c9099547b5ee37ee750,

title = "Training a neural network based on unreliable human annotation of medical images",

abstract = "Building classification models from clinical data often requires human experts for example labeling. However, it is difficult to obtain a perfect set of labels due to the complexity of the medical data and the large variability between experts. In this study we present a neural-network training strategy that is more robust to unreliable labeling by explicitly modeling the label noise as part of the network architecture. Our method is demonstrated on breast microcalcifications classification into benign and malignant categories, given multi-view mammograms. We show that the proposed training procedure outperforms standard training methods that ignore the existence of label noise.",

keywords = "Deep-learning, Mammography, Microcalcifications, Noisy-labels, Robust training",

author = "Yair Dgani and Hayit Greenspan and Jacob Goldberger",

note = "Publisher Copyright: {\textcopyright} 2018 IEEE.; 15th IEEE International Symposium on Biomedical Imaging, ISBI 2018 ; Conference date: 04-04-2018 Through 07-04-2018",

year = "2018",

month = may,

day = "23",

doi = "10.1109/ISBI.2018.8363518",

language = "אנגלית",

series = "Proceedings - International Symposium on Biomedical Imaging",

publisher = "IEEE Computer Society",

pages = "39--42",

booktitle = "2018 IEEE 15th International Symposium on Biomedical Imaging, ISBI 2018",

address = "ארצות הברית",

}

Dgani, Y, Greenspan, H & Goldberger, J 2018, Training a neural network based on unreliable human annotation of medical images. in 2018 IEEE 15th International Symposium on Biomedical Imaging, ISBI 2018. Proceedings - International Symposium on Biomedical Imaging, vol. 2018-April, IEEE Computer Society, pp. 39-42, 15th IEEE International Symposium on Biomedical Imaging, ISBI 2018, Washington, United States, 4/04/18. https://doi.org/10.1109/ISBI.2018.8363518

Training a neural network based on unreliable human annotation of medical images. / Dgani, Yair; Greenspan, Hayit; Goldberger, Jacob.
2018 IEEE 15th International Symposium on Biomedical Imaging, ISBI 2018. IEEE Computer Society, 2018. p. 39-42 (Proceedings - International Symposium on Biomedical Imaging; Vol. 2018-April).

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

TY - GEN

T1 - Training a neural network based on unreliable human annotation of medical images

AU - Dgani, Yair

AU - Greenspan, Hayit

AU - Goldberger, Jacob

PY - 2018/5/23

Y1 - 2018/5/23

N2 - Building classification models from clinical data often requires human experts for example labeling. However, it is difficult to obtain a perfect set of labels due to the complexity of the medical data and the large variability between experts. In this study we present a neural-network training strategy that is more robust to unreliable labeling by explicitly modeling the label noise as part of the network architecture. Our method is demonstrated on breast microcalcifications classification into benign and malignant categories, given multi-view mammograms. We show that the proposed training procedure outperforms standard training methods that ignore the existence of label noise.

AB - Building classification models from clinical data often requires human experts for example labeling. However, it is difficult to obtain a perfect set of labels due to the complexity of the medical data and the large variability between experts. In this study we present a neural-network training strategy that is more robust to unreliable labeling by explicitly modeling the label noise as part of the network architecture. Our method is demonstrated on breast microcalcifications classification into benign and malignant categories, given multi-view mammograms. We show that the proposed training procedure outperforms standard training methods that ignore the existence of label noise.

KW - Deep-learning

KW - Mammography

KW - Microcalcifications

KW - Noisy-labels

KW - Robust training

UR - http://www.scopus.com/inward/record.url?scp=85048129467&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2018.8363518

DO - 10.1109/ISBI.2018.8363518

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85048129467

T3 - Proceedings - International Symposium on Biomedical Imaging

SP - 39

EP - 42

BT - 2018 IEEE 15th International Symposium on Biomedical Imaging, ISBI 2018

PB - IEEE Computer Society

T2 - 15th IEEE International Symposium on Biomedical Imaging, ISBI 2018

Y2 - 4 April 2018 through 7 April 2018

ER -

Training a neural network based on unreliable human annotation of medical images

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this