There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net .
Beschreibung
U-Net: Convolutional Networks for Biomedical Image Segmentation | SpringerLink
Medical Image Computing and Computer-Assisted Intervention -- MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III
%0 Book Section
%1 journals/corr/RonnebergerFB15
%A Ronneberger, Olaf
%A Fischer, Philipp
%A Brox, Thomas
%B Medical Image Computing and Computer-Assisted Intervention -- MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III
%C Cham
%D 2015
%E Navab, Nassir
%E Hornegger, Joachim
%E Wells, William M.
%E Frangi, Alejandro F.
%I Springer International Publishing
%K DAS_2018_1 DNN
%P 234--241
%R 10.1007/978-3-319-24574-4_28
%T U-Net: Convolutional Networks for Biomedical Image Segmentation
%U https://doi.org/10.1007/978-3-319-24574-4_28
%X There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net .
%@ 978-3-319-24574-4
@inbook{journals/corr/RonnebergerFB15,
abstract = {There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net .},
added-at = {2017-11-09T09:53:44.000+0100},
address = {Cham},
author = {Ronneberger, Olaf and Fischer, Philipp and Brox, Thomas},
biburl = {https://www.bibsonomy.org/bibtex/26622be497b411a27594657a531a625b4/chwick},
booktitle = {Medical Image Computing and Computer-Assisted Intervention -- MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III},
description = {U-Net: Convolutional Networks for Biomedical Image Segmentation | SpringerLink},
doi = {10.1007/978-3-319-24574-4_28},
editor = {Navab, Nassir and Hornegger, Joachim and Wells, William M. and Frangi, Alejandro F.},
interhash = {9158de16b2caff7458df054dc6fc2748},
intrahash = {6622be497b411a27594657a531a625b4},
isbn = {978-3-319-24574-4},
keywords = {DAS_2018_1 DNN},
pages = {234--241},
publisher = {Springer International Publishing},
timestamp = {2017-11-09T09:53:44.000+0100},
title = {U-Net: Convolutional Networks for Biomedical Image Segmentation},
url = {https://doi.org/10.1007/978-3-319-24574-4_28},
year = 2015
}