[1]谭璐露,冯前进.基于深度卷积和三向注意力感知的胰腺分割算法[J].中国医学物理学杂志,2025,42(1):37-42.[doi:DOI:10.3969/j.issn.1005-202X.2025.01.006]
　TAN Lulu,FENG Qianjin,Pancreas segmentation algorithm based on depth-wise convolution and tri-orientated spatial attention[J].Chinese Journal of Medical Physics,2025,42(1):37-42.[doi:DOI:10.3969/j.issn.1005-202X.2025.01.006]
点击复制

基于深度卷积和三向注意力感知的胰腺分割算法()
分享到：

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
42
期数:
2025年第1期
页码:
37-42
栏目:
医学影像物理
出版日期:
2025-01-19

文章信息/Info

Title:
Pancreas segmentation algorithm based on depth-wise convolution and tri-orientated spatial attention
文章编号:
1005-202X（2025）01-0037-06
作者:
谭璐露1; 冯前进1; 2
1.南方医科大学生物医学工程学院， 广东 广州 510515； 2.广东省医学图像处理重点实验室， 广东 广州 510515
Author(s):
TAN Lulu1;  FENG Qianjin1;  2
1. School of Biomedical Engineering, Southern Medical University, Guangzhou 510515, China 2. Guangdong Provincial Key Laboratory of Medical Image Processing, Guangzhou 510515, China
关键词:
胰腺; 深度卷积; 三向注意力; 级联网络
Keywords:
Keywords: pancreas depth-wise convolution tri-orientated spatial attention cascaded network
分类号:
R318
DOI:
DOI:10.3969/j.issn.1005-202X.2025.01.006
文献标志码:
A
摘要:
针对胰腺分割任务中因体积较小且解剖结构复杂带来的挑战，提出一种级联的3D胰腺分割网络（CPS-Net）。CPS-Net由两部分组成：第一部分采用ResUNet快速定位胰腺区域，第二部分使用融合深度卷积（DCB）和三向注意力感知模块（ToSA）的网络来细化分割结果。DCB通过逐层提取多尺度特征，显著增强胰腺与周围组织之间的区分能力。而ToSA则结合轴向注意力、平面注意力和窗口注意力机制，全面捕捉胰腺在复杂背景中的细节结构。CPS-Net在NIH公开数据集上的Dice相似性系数、阳性预测值、敏感性和Hausdorff距离指标分别达到（87.42±1.58）%、（87.42±3.52）%、（87.74±4.58）%和（0.22±0.08） mm。实验结果表明，CPS-Net表现优于当前主流分割网络，显著提升胰腺分割精度。 【关键词】胰腺；深度卷积；三向注意力；级联网络
Abstract:
Abstract: A cascaded 3D pancreas segmentation network (CPS-Net) is proposed to address the challenges in pancreas segmentation caused by its small size and complex anatomical structure. CPS-Net is composed of two components: the first part utilizes ResUNet to quickly localize the pancreas region, while the second part uses a network that fuses depth-wise convolution block and tri-orientated spatial attention module to refine the segmentation results. Specifically, depth-wise convolution block significantly enhances the differentiation between the pancreas and surrounding tissues by extracting multi-scale features layer by layer, while tri-orientated spatial attention module combines axial attention, planar attention and window attention mechanisms to comprehensively capture the detailed structure of the pancreas in a complex background. CPS-Net achieved Dice similarity coefficient, positive predictive value, sensitivity, and Hausdorff distance of 87.42%±1.58%, 87.42%±3.52%, 87.74%±4.58%, and (0.22±0.08) mm, respectively, on the NIH public dataset, demonstrating its higher pancreas segmentation accuracy and superior performance compared with the current state-of-the-art segmentation networks.

相似文献/References:

[1]陆敏达,张财源,侯亮,等. 儿童胰腺实性假乳头状瘤多排螺旋CT表现[J].中国医学物理学杂志,2018,35(8):909.[doi:DOI:10.3969/j.issn.1005-202X.2018.08.009]
　LU Minda,ZHANG Caiyuan,HOU Liang,et al. Multidetector computed tomography features of solid pseudopapillary tumor of pancreas in children[J].Chinese Journal of Medical Physics,2018,35(1):909.[doi:DOI:10.3969/j.issn.1005-202X.2018.08.009]
[2]曹洋森,朱晓斐,韩妙飞,等.基于级联式深度网络模型的胃及胰腺自动分割研究[J].中国医学物理学杂志,2021,38(8):971.[doi:DOI:10.3969/j.issn.1005-202X.2021.08.010]
　CAO Yangsen,ZHU Xiaofei,HAN Miaofei,et al.Automatic segmentation of the stomach and pancreas using cascaded deep convolutional neural network[J].Chinese Journal of Medical Physics,2021,38(1):971.[doi:DOI:10.3969/j.issn.1005-202X.2021.08.010]

备注/Memo

备注/Memo:
【收稿日期】2024-10-23 【基金项目】国家自然科学基金（62471214） 【作者简介】谭璐露，硕士研究生，研究方向：医学人工智能，E-mail: t13763964685@163.com 【通信作者】冯前进，博士，教授，博士生导师，研究方向：医学图像分析，E-mail: Fengqj99@smu.edu.cn
更新日期/Last Update: 2025-01-19