TY - GEN
T1 - Unet++
T2 - 4th International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2018 and 8th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2018 Held in Conjunction with MICCAI 2018
AU - Zhou, Zongwei
AU - Rahman Siddiquee, Md Mahfuzur
AU - Tajbakhsh, Nima
AU - Liang, Jianming
N1 - Funding Information:
Acknowledgments. This research has been supported partially by NIH under Award Number R01HL128785, by ASU and Mayo Clinic through a Seed Grant and an Innovation Grant. The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH.
Publisher Copyright:
© Springer Nature Switzerland AG 2018.
PY - 2018
Y1 - 2018
N2 - In this paper, we present UNet++, a new, more powerful architecture for medical image segmentation. Our architecture is essentially a deeply-supervised encoder-decoder network where the encoder and decoder sub-networks are connected through a series of nested, dense skip pathways. The re-designed skip pathways aim at reducing the semantic gap between the feature maps of the encoder and decoder sub-networks. We argue that the optimizer would deal with an easier learning task when the feature maps from the decoder and encoder networks are semantically similar. We have evaluated UNet++ in comparison with U-Net and wide U-Net architectures across multiple medical image segmentation tasks: nodule segmentation in the low-dose CT scans of chest, nuclei segmentation in the microscopy images, liver segmentation in abdominal CT scans, and polyp segmentation in colonoscopy videos. Our experiments demonstrate that UNet++ with deep supervision achieves an average IoU gain of 3.9 and 3.4 points over U-Net and wide U-Net, respectively.
AB - In this paper, we present UNet++, a new, more powerful architecture for medical image segmentation. Our architecture is essentially a deeply-supervised encoder-decoder network where the encoder and decoder sub-networks are connected through a series of nested, dense skip pathways. The re-designed skip pathways aim at reducing the semantic gap between the feature maps of the encoder and decoder sub-networks. We argue that the optimizer would deal with an easier learning task when the feature maps from the decoder and encoder networks are semantically similar. We have evaluated UNet++ in comparison with U-Net and wide U-Net architectures across multiple medical image segmentation tasks: nodule segmentation in the low-dose CT scans of chest, nuclei segmentation in the microscopy images, liver segmentation in abdominal CT scans, and polyp segmentation in colonoscopy videos. Our experiments demonstrate that UNet++ with deep supervision achieves an average IoU gain of 3.9 and 3.4 points over U-Net and wide U-Net, respectively.
UR - http://www.scopus.com/inward/record.url?scp=85054551189&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85054551189&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-00889-5_1
DO - 10.1007/978-3-030-00889-5_1
M3 - Conference contribution
AN - SCOPUS:85054551189
SN - 9783030008888
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 3
EP - 11
BT - Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 4th International Workshop, DLMIA 2018 and 8th International Workshop, ML-CDS 2018 Held in Conjunction with MICCAI 2018
A2 - Maier-Hein, Lena
A2 - Syeda-Mahmood, Tanveer
A2 - Taylor, Zeike
A2 - Lu, Zhi
A2 - Stoyanov, Danail
A2 - Madabhushi, Anant
A2 - Tavares, João Manuel R.S.
A2 - Nascimento, Jacinto C.
A2 - Moradi, Mehdi
A2 - Martel, Anne
A2 - Papa, Joao Paulo
A2 - Conjeti, Sailesh
A2 - Belagiannis, Vasileios
A2 - Greenspan, Hayit
A2 - Carneiro, Gustavo
A2 - Bradley, Andrew
PB - Springer Verlag
Y2 - 20 September 2018 through 20 September 2018
ER -