Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss

Yu-Cheng Liu; Daniel Stanley Tan; Jyh-Cheng Chen; Wen-Huang Cheng; Kai-Lung Hua

doi:10.1109/icip.2019.8803471

Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss

Yu-Cheng Liu, Daniel Stanley Tan, Jyh-Cheng Chen, Wen-Huang Cheng, Kai-Lung Hua

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

Abstract

We propose a novel network architecture called Residual Attention U-Net (ResAttU-Net) for segmenting hepatic lesions. Our model incorporates residual blocks that can extract more complex features as compared with traditional convolutional layers combined with a skip-connection attention module that learns to focus on the relevant features for the task of hepatic lesions segmentation. Moreover, we train our model using an adaptive weighted dice loss that prioritizes the pixels of the tumor class over the pixels of the background class. We evaluate our model on the MICCAI Liver Tumor Segmentation (LiTS) benchmark dataset. Our experimental results show that our method significantly improves upon several state-of-the-art baselines for hepatic lesion or liver tumor segmentation.

Original language	English
Title of host publication	2019 IEEE International Conference on Image Processing (ICIP)
Subtitle of host publication	Proceedings
Publisher	IEEE
Pages	3322-3326
Number of pages	5
ISBN (Electronic)	978-1-5386-6249-6
ISBN (Print)	978-1-5386-6250-2
DOIs	https://doi.org/10.1109/icip.2019.8803471
Publication status	Published - Sept 2019
Externally published	Yes
Event	IEEE International Conference on Image Processing - Taipei, Taiwan, Province of China Duration: 22 Sept 2019 → 25 Sept 2019 https://ieeexplore.ieee.org/xpl/conhome/8791230/proceeding

Conference

Conference	IEEE International Conference on Image Processing
Abbreviated title	ICIP 2019
Country/Territory	Taiwan, Province of China
City	Taipei
Period	22/09/19 → 25/09/19
Internet address	https://ieeexplore.ieee.org/xpl/conhome/8791230/proceeding

Keywords

CT image segmentation
attention module
hepatic lesion factor
residual block

Access to Document

10.1109/icip.2019.8803471

Cite this

@inproceedings{f6b2a0e4e3b04f8880c0dc823a90a421,

title = "Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss",

abstract = "We propose a novel network architecture called Residual Attention U-Net (ResAttU-Net) for segmenting hepatic lesions. Our model incorporates residual blocks that can extract more complex features as compared with traditional convolutional layers combined with a skip-connection attention module that learns to focus on the relevant features for the task of hepatic lesions segmentation. Moreover, we train our model using an adaptive weighted dice loss that prioritizes the pixels of the tumor class over the pixels of the background class. We evaluate our model on the MICCAI Liver Tumor Segmentation (LiTS) benchmark dataset. Our experimental results show that our method significantly improves upon several state-of-the-art baselines for hepatic lesion or liver tumor segmentation.",

keywords = "CT image segmentation, attention module, hepatic lesion factor, residual block",

author = "Yu-Cheng Liu and Tan, {Daniel Stanley} and Jyh-Cheng Chen and Wen-Huang Cheng and Kai-Lung Hua",

year = "2019",

month = sep,

doi = "10.1109/icip.2019.8803471",

language = "English",

isbn = "978-1-5386-6250-2",

pages = "3322--3326",

booktitle = "2019 IEEE International Conference on Image Processing (ICIP)",

publisher = "IEEE",

address = "United States",

note = "IEEE International Conference on Image Processing, ICIP 2019 ; Conference date: 22-09-2019 Through 25-09-2019",

url = "https://ieeexplore.ieee.org/xpl/conhome/8791230/proceeding",

}

Liu, Y-C, Tan, DS, Chen, J-C, Cheng, W-H & Hua, K-L 2019, Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss. in 2019 IEEE International Conference on Image Processing (ICIP): Proceedings. IEEE, pp. 3322-3326, IEEE International Conference on Image Processing, Taipei, Taiwan, Province of China, 22/09/19. https://doi.org/10.1109/icip.2019.8803471

Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss. / Liu, Yu-Cheng; Tan, Daniel Stanley; Chen, Jyh-Cheng et al.
2019 IEEE International Conference on Image Processing (ICIP): Proceedings. IEEE, 2019. p. 3322-3326.

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

TY - GEN

T1 - Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss

AU - Liu, Yu-Cheng

AU - Tan, Daniel Stanley

AU - Chen, Jyh-Cheng

AU - Cheng, Wen-Huang

AU - Hua, Kai-Lung

PY - 2019/9

Y1 - 2019/9

N2 - We propose a novel network architecture called Residual Attention U-Net (ResAttU-Net) for segmenting hepatic lesions. Our model incorporates residual blocks that can extract more complex features as compared with traditional convolutional layers combined with a skip-connection attention module that learns to focus on the relevant features for the task of hepatic lesions segmentation. Moreover, we train our model using an adaptive weighted dice loss that prioritizes the pixels of the tumor class over the pixels of the background class. We evaluate our model on the MICCAI Liver Tumor Segmentation (LiTS) benchmark dataset. Our experimental results show that our method significantly improves upon several state-of-the-art baselines for hepatic lesion or liver tumor segmentation.

AB - We propose a novel network architecture called Residual Attention U-Net (ResAttU-Net) for segmenting hepatic lesions. Our model incorporates residual blocks that can extract more complex features as compared with traditional convolutional layers combined with a skip-connection attention module that learns to focus on the relevant features for the task of hepatic lesions segmentation. Moreover, we train our model using an adaptive weighted dice loss that prioritizes the pixels of the tumor class over the pixels of the background class. We evaluate our model on the MICCAI Liver Tumor Segmentation (LiTS) benchmark dataset. Our experimental results show that our method significantly improves upon several state-of-the-art baselines for hepatic lesion or liver tumor segmentation.

KW - CT image segmentation

KW - attention module

KW - hepatic lesion factor

KW - residual block

U2 - 10.1109/icip.2019.8803471

DO - 10.1109/icip.2019.8803471

M3 - Conference Article in proceeding

SN - 978-1-5386-6250-2

SP - 3322

EP - 3326

BT - 2019 IEEE International Conference on Image Processing (ICIP)

PB - IEEE

T2 - IEEE International Conference on Image Processing

Y2 - 22 September 2019 through 25 September 2019

ER -

Segmenting Hepatic Lesions Using Residual Attention U-Net with an Adaptive Weighted Dice Loss

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this