Convolutional neural networks (CNNs) have become popular in medical image segmentation. The widely used deep CNNs are customized to extract multiple representative features for two-dimensional (2D) data, generally called 2D networks. However, 2D networks are inefficient in extracting three-dimensional (3D) spatial features from volumetric images. Although most 2D segmentation networks can be extended to 3D networks, the naively extended 3D methods are resource-intensive. In this paper, we propose an efficient and accurate network for fully automatic 3D segmentation. Specifically, we designed a 3D multiple-contextual extractor to capture rich global contextual dependencies from different feature levels. Then we leveraged an ROI-estimation strategy to crop the ROI bounding box. Meanwhile, we used a 3D ROI-attention module to improve the accuracy of in-region segmentation in the decoder path. Moreover, we used a hybrid Dice loss function to address the issues of class imbalance and blurry contour in medical images. By incorporating the above strategies, we realized a practical end-to-end 3D medical image segmentation with high efficiency and accuracy. To validate the 3D segmentation performance of our proposed method, we conducted extensive experiments on two datasets and demonstrated favorable results over the state-of-the-art methods.
He LI
Ritsumeikan University
Yutaro IWAMOTO
Ritsumeikan University
Xianhua HAN
Yamaguchi University
Lanfen LIN
Zhejiang University
Akira FURUKAWA
Tokyo Metropolitan University
Shuzo KANASAKI
Koseikai Takeda Hospital
Yen-Wei CHEN
Ritsumeikan University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
He LI, Yutaro IWAMOTO, Xianhua HAN, Lanfen LIN, Akira FURUKAWA, Shuzo KANASAKI, Yen-Wei CHEN, "3D Multiple-Contextual ROI-Attention Network for Efficient and Accurate Volumetric Medical Image Segmentation" in IEICE TRANSACTIONS on Information,
vol. E106-D, no. 5, pp. 1027-1037, May 2023, doi: 10.1587/transinf.2022EDP7193.
Abstract: Convolutional neural networks (CNNs) have become popular in medical image segmentation. The widely used deep CNNs are customized to extract multiple representative features for two-dimensional (2D) data, generally called 2D networks. However, 2D networks are inefficient in extracting three-dimensional (3D) spatial features from volumetric images. Although most 2D segmentation networks can be extended to 3D networks, the naively extended 3D methods are resource-intensive. In this paper, we propose an efficient and accurate network for fully automatic 3D segmentation. Specifically, we designed a 3D multiple-contextual extractor to capture rich global contextual dependencies from different feature levels. Then we leveraged an ROI-estimation strategy to crop the ROI bounding box. Meanwhile, we used a 3D ROI-attention module to improve the accuracy of in-region segmentation in the decoder path. Moreover, we used a hybrid Dice loss function to address the issues of class imbalance and blurry contour in medical images. By incorporating the above strategies, we realized a practical end-to-end 3D medical image segmentation with high efficiency and accuracy. To validate the 3D segmentation performance of our proposed method, we conducted extensive experiments on two datasets and demonstrated favorable results over the state-of-the-art methods.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2022EDP7193/_p
Copy
@ARTICLE{e106-d_5_1027,
author={He LI, Yutaro IWAMOTO, Xianhua HAN, Lanfen LIN, Akira FURUKAWA, Shuzo KANASAKI, Yen-Wei CHEN, },
journal={IEICE TRANSACTIONS on Information},
title={3D Multiple-Contextual ROI-Attention Network for Efficient and Accurate Volumetric Medical Image Segmentation},
year={2023},
volume={E106-D},
number={5},
pages={1027-1037},
abstract={Convolutional neural networks (CNNs) have become popular in medical image segmentation. The widely used deep CNNs are customized to extract multiple representative features for two-dimensional (2D) data, generally called 2D networks. However, 2D networks are inefficient in extracting three-dimensional (3D) spatial features from volumetric images. Although most 2D segmentation networks can be extended to 3D networks, the naively extended 3D methods are resource-intensive. In this paper, we propose an efficient and accurate network for fully automatic 3D segmentation. Specifically, we designed a 3D multiple-contextual extractor to capture rich global contextual dependencies from different feature levels. Then we leveraged an ROI-estimation strategy to crop the ROI bounding box. Meanwhile, we used a 3D ROI-attention module to improve the accuracy of in-region segmentation in the decoder path. Moreover, we used a hybrid Dice loss function to address the issues of class imbalance and blurry contour in medical images. By incorporating the above strategies, we realized a practical end-to-end 3D medical image segmentation with high efficiency and accuracy. To validate the 3D segmentation performance of our proposed method, we conducted extensive experiments on two datasets and demonstrated favorable results over the state-of-the-art methods.},
keywords={},
doi={10.1587/transinf.2022EDP7193},
ISSN={1745-1361},
month={May},}
Copy
TY - JOUR
TI - 3D Multiple-Contextual ROI-Attention Network for Efficient and Accurate Volumetric Medical Image Segmentation
T2 - IEICE TRANSACTIONS on Information
SP - 1027
EP - 1037
AU - He LI
AU - Yutaro IWAMOTO
AU - Xianhua HAN
AU - Lanfen LIN
AU - Akira FURUKAWA
AU - Shuzo KANASAKI
AU - Yen-Wei CHEN
PY - 2023
DO - 10.1587/transinf.2022EDP7193
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E106-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2023
AB - Convolutional neural networks (CNNs) have become popular in medical image segmentation. The widely used deep CNNs are customized to extract multiple representative features for two-dimensional (2D) data, generally called 2D networks. However, 2D networks are inefficient in extracting three-dimensional (3D) spatial features from volumetric images. Although most 2D segmentation networks can be extended to 3D networks, the naively extended 3D methods are resource-intensive. In this paper, we propose an efficient and accurate network for fully automatic 3D segmentation. Specifically, we designed a 3D multiple-contextual extractor to capture rich global contextual dependencies from different feature levels. Then we leveraged an ROI-estimation strategy to crop the ROI bounding box. Meanwhile, we used a 3D ROI-attention module to improve the accuracy of in-region segmentation in the decoder path. Moreover, we used a hybrid Dice loss function to address the issues of class imbalance and blurry contour in medical images. By incorporating the above strategies, we realized a practical end-to-end 3D medical image segmentation with high efficiency and accuracy. To validate the 3D segmentation performance of our proposed method, we conducted extensive experiments on two datasets and demonstrated favorable results over the state-of-the-art methods.
ER -