Exploiting the potential of unlabeled endoscopic video data with self-supervised learning
Purpose Surgical data science is a new research field that aims to observe all aspects of the patient treatment process in order to provide the right assistance at the right time. Due to the breakthrough successes of deep learning-based solutions for automatic image annotation, the availability of r...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
27 April 2018
|
| In: |
International journal of computer assisted radiology and surgery
Year: 2018, Volume: 13, Issue: 6, Pages: 925-933 |
| ISSN: | 1861-6429 |
| DOI: | 10.1007/s11548-018-1772-0 |
| Online Access: | Verlag, Volltext: https://doi.org/10.1007/s11548-018-1772-0
|
| Author Notes: | Tobias Ross, David Zimmerer, Anant Vemuri, Fabian Isensee, Manuel Wiesenfarth, Sebastian Bodenstedt, Fabian Both, Philip Kessler, Martin Wagner, Beat Müller, Hannes Kenngott, Stefanie Speidel, Annette Kopp-Schneider, Klaus Maier-Hein, Lena Maier-Hein |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1663271186 | ||
| 003 | DE-627 | ||
| 005 | 20220816134625.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 190415s2018 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1007/s11548-018-1772-0 |2 doi | |
| 035 | |a (DE-627)1663271186 | ||
| 035 | |a (DE-599)KXP1663271186 | ||
| 035 | |a (OCoLC)1341209678 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 33 |2 sdnb | ||
| 100 | 1 | |a Roß, Tobias |d 1990- |e VerfasserIn |0 (DE-588)118364955X |0 (DE-627)1663271143 |4 aut | |
| 245 | 1 | 0 | |a Exploiting the potential of unlabeled endoscopic video data with self-supervised learning |c Tobias Ross, David Zimmerer, Anant Vemuri, Fabian Isensee, Manuel Wiesenfarth, Sebastian Bodenstedt, Fabian Both, Philip Kessler, Martin Wagner, Beat Müller, Hannes Kenngott, Stefanie Speidel, Annette Kopp-Schneider, Klaus Maier-Hein, Lena Maier-Hein |
| 264 | 1 | |c 27 April 2018 | |
| 300 | |a 9 | ||
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Gesehen am 15.04.2019 | ||
| 520 | |a Purpose Surgical data science is a new research field that aims to observe all aspects of the patient treatment process in order to provide the right assistance at the right time. Due to the breakthrough successes of deep learning-based solutions for automatic image annotation, the availability of reference annotations for algorithm training is becoming a major bottleneck in the field. The purpose of this paper was to investigate the concept of self-supervised learning to address this issue. Methods Our approach is guided by the hypothesis that unlabeled video data can be used to learn a representation of the target domain that boosts the performance of state-of-the-art machine learning algorithms when used for pre-training. Core of the method is an auxiliary task based on raw endoscopic video data of the target domain that is used to initialize the convolutional neural network (CNN) for the target task. In this paper, we propose the re-colorization of medical images with a conditional generative adversarial network (cGAN)-based architecture as auxiliary task. A variant of the method involves a second pre-training step based on labeled data for the target task from a related domain. We validate both variants using medical instrument segmentation as target task. Results The proposed approach can be used to radically reduce the manual annotation effort involved in training CNNs. Compared to the baseline approach of generating annotated data from scratch, our method decreases exploratively the number of labeled images by up to 75% without sacrificing performance. Our method also outperforms alternative methods for CNN pre-training, such as pre-training on publicly available non-medical (COCO) or medical data (MICCAI EndoVis2017 challenge) using the target task (in this instance: segmentation). Conclusion As it makes efficient use of available (non-)public and (un-)labeled data, the approach has the potential to become a valuable tool for CNN (pre-)training. | ||
| 650 | 4 | |a Computer vision | |
| 650 | 4 | |a Endoscopic image processing | |
| 650 | 4 | |a Endoscopic instrument segmentation | |
| 650 | 4 | |a Self-supervised learning | |
| 650 | 4 | |a Transfer learning | |
| 700 | 1 | |a Wagner, Martin |d 1988- |e VerfasserIn |0 (DE-588)1060231980 |0 (DE-627)799324337 |0 (DE-576)416324282 |4 aut | |
| 700 | 1 | |a Müller, Beat P. |d 1971- |e VerfasserIn |0 (DE-588)14066209X |0 (DE-627)70374819X |0 (DE-576)317992287 |4 aut | |
| 700 | 1 | |a Kenngott, Hannes Götz |d 1979- |e VerfasserIn |0 (DE-588)141469994 |0 (DE-627)62780117X |0 (DE-576)324023065 |4 aut | |
| 700 | 1 | |a Maier-Hein, Lena |d 1980- |e VerfasserIn |0 (DE-588)1075029252 |0 (DE-627)832869899 |0 (DE-576)190090804 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t International journal of computer assisted radiology and surgery |d Berlin : Springer, 2006 |g 13(2018), 6, Seite 925-933 |h Online-Ressource |w (DE-627)512299250 |w (DE-600)2235881-X |w (DE-576)283119322 |x 1861-6429 |7 nnas |a Exploiting the potential of unlabeled endoscopic video data with self-supervised learning |
| 773 | 1 | 8 | |g volume:13 |g year:2018 |g number:6 |g pages:925-933 |g extent:9 |a Exploiting the potential of unlabeled endoscopic video data with self-supervised learning |
| 856 | 4 | 0 | |u https://doi.org/10.1007/s11548-018-1772-0 |x Verlag |x Resolving-System |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20190415 | ||
| 993 | |a Article | ||
| 994 | |a 2018 | ||
| 998 | |g 1075029252 |a Maier-Hein, Lena |m 1075029252:Maier-Hein, Lena |d 50000 |e 50000PM1075029252 |k 0/50000/ |p 15 |y j | ||
| 998 | |g 141469994 |a Kenngott, Hannes Götz |m 141469994:Kenngott, Hannes Götz |d 910000 |d 910200 |e 910000PK141469994 |e 910200PK141469994 |k 0/910000/ |k 1/910000/910200/ |p 11 | ||
| 998 | |g 14066209X |a Müller, Beat P. |m 14066209X:Müller, Beat P. |d 910000 |d 910200 |e 910000PM14066209X |e 910200PM14066209X |k 0/910000/ |k 1/910000/910200/ |p 10 | ||
| 998 | |g 1060231980 |a Wagner, Martin |m 1060231980:Wagner, Martin |d 910000 |d 910200 |e 910000PW1060231980 |e 910200PW1060231980 |k 0/910000/ |k 1/910000/910200/ |p 9 | ||
| 999 | |a KXP-PPN1663271186 |e 3422496297 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"language":["eng"],"origin":[{"dateIssuedKey":"2018","dateIssuedDisp":"27 April 2018"}],"type":{"bibl":"article-journal","media":"Online-Ressource"},"name":{"displayForm":["Tobias Ross, David Zimmerer, Anant Vemuri, Fabian Isensee, Manuel Wiesenfarth, Sebastian Bodenstedt, Fabian Both, Philip Kessler, Martin Wagner, Beat Müller, Hannes Kenngott, Stefanie Speidel, Annette Kopp-Schneider, Klaus Maier-Hein, Lena Maier-Hein"]},"person":[{"roleDisplay":"VerfasserIn","family":"Roß","role":"aut","display":"Roß, Tobias","given":"Tobias"},{"family":"Wagner","roleDisplay":"VerfasserIn","given":"Martin","role":"aut","display":"Wagner, Martin"},{"roleDisplay":"VerfasserIn","family":"Müller","given":"Beat P.","display":"Müller, Beat P.","role":"aut"},{"roleDisplay":"VerfasserIn","family":"Kenngott","given":"Hannes Götz","role":"aut","display":"Kenngott, Hannes Götz"},{"role":"aut","display":"Maier-Hein, Lena","given":"Lena","family":"Maier-Hein","roleDisplay":"VerfasserIn"}],"physDesc":[{"extent":"9 S."}],"relHost":[{"origin":[{"dateIssuedDisp":"2006-","publisherPlace":"Berlin ; Heidelberg [u.a.]","dateIssuedKey":"2006","publisher":"Springer"}],"type":{"media":"Online-Ressource","bibl":"periodical"},"language":["eng"],"physDesc":[{"extent":"Online-Ressource"}],"part":{"text":"13(2018), 6, Seite 925-933","volume":"13","issue":"6","extent":"9","year":"2018","pages":"925-933"},"id":{"issn":["1861-6429"],"zdb":["2235881-X"],"eki":["512299250"]},"title":[{"subtitle":"a journal for interdisciplinary research, development and applications of image guided diagnosis and therapy","title_sort":"International journal of computer assisted radiology and surgery","title":"International journal of computer assisted radiology and surgery"}],"note":["Gesehen am 26.02.14","Fortsetzung der Druck-Ausgabe"],"disp":"Exploiting the potential of unlabeled endoscopic video data with self-supervised learningInternational journal of computer assisted radiology and surgery","recId":"512299250","pubHistory":["1.2006/07 -"]}],"id":{"doi":["10.1007/s11548-018-1772-0"],"eki":["1663271186"]},"title":[{"title":"Exploiting the potential of unlabeled endoscopic video data with self-supervised learning","title_sort":"Exploiting the potential of unlabeled endoscopic video data with self-supervised learning"}],"recId":"1663271186","note":["Gesehen am 15.04.2019"]} | ||
| SRT | |a ROSSTOBIASEXPLOITING2720 | ||