摘 要：现有的医学图像器官分割方法不能很好地依肝脏形状、位置及大小的变化而进行适当的分割，当肝脏形态变化明显时，不能准确地将肝脏分割出来。鉴于此，文章在传统 U-Net 网络中加入了全局注意力模块，通过通道注意力和自我注意力增强了对肝脏的特征提取；并在自动分割的基础上进行了人机协同操作，对分割不好的部分增加数据量，有效提高了分割准确率。该模型在 MIOU 和 MPA 指标上分别达到了 86.71%、92.58%。
关键词：医学影像；人机协同；器官分割；U-Net 网络
DOI:10.19850/j.cnki.2096-4706.2023.06.014
中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2023）06-0054-04
Automatic Liver Segmentation of CT Images Based on FU-Net Network of Human-Computer Cooperation
WANG Jiaqi, ZHANG Guangyuan, LI Kefeng
(Shandong Jiaotong University, Jinan 250357, China)
Abstract: The existing organ segmentation methods in medical images can not segment properly according to the changes of liver shape, position and size. When the liver shape changes obviously, the liver can not be accurately segmented. In view of this, this paper adds a global attention module to the traditional U-Net network, which enhances the feature extraction of liver through channel attention and self attention. On the basis of automatic segmentation, human-computer cooperation is carried out to increase the amount of data for the bad part of segmentation and effectively improve the accuracy of segmentation. The model reaches 86.71% and 92.58% respectively in MIOU and MPA indicators.
medical image; human-computer cooperation; organ segmentation; U-Net network
参考文献：
[1] LONG J，SHELHAMER E，D A R R E L L T. F u l l y Convolutional Networks for Semantic Segmentation [J/OL]. arXiv:1411.4038 [cs.CV].[2022-10-18].https://arxiv.org/abs/1411.4038.
[2] RONNEBERGER O，FISCHER P BROX T. U-net： Convolutional networks for biomedical image segmentation [EB/OL]. [2022-10-19].https://www.doc88.com/p-9909769280204.html.
[3] BADRINARAYANAN V，KENDALL A，CIPOLLA R. Segnet：A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（12）：2481-2495.
[4] ZHAO H S，SHI J P，QI X J，et al. Pyramid scene parsing network [C]//2017 IEEE conference on computer vision and pattern recognition.Honolulu：IEEE，2017：6230-6239.
[5] ZHANG H，DANA K，SHI J P，et al. Context Encoding for Semantic Segmentation [C]//2018 IEEE/CVF conference on Computer Vision and Pattern Recognition.Salt Lake City：IEEE，2018：7151-7160.
[6] ZHOU D Q，YU Z D，XIE E Z，et al. Understanding the robustness in vision transformers [J/OL].arXiv:2204.12451 [cs.CV]. [2022-10-18].https://arxiv.org/abs/2204.12451v4.
[7] VASWANI A，SHAZEER N，PARMAR N，et al. Attention is all You Need [J/OL].arXiv：1706.03762 [cs.CL].[2022-10-18].https:// arxiv.org/abs/1706.03762v2.
[8] HU J，SHEN L，ALBANIE S. Squeeze-and-Excitation Networks [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2020，42（8）：2011-2023.
作者简介：王佳琪（1996—），女，汉族，山东烟台人，硕士在读，研究方向：电子电气。