Abstract：There are many challenges in current lung trachea segmentation including complex grayscale distribution of CT images, segmentation target pixel approximation, easy to cause over-segmentation, fewer lung trachea pixels, and difficult to get more target features, therefore, fine lung trachea is easy to be ignored. In this paper, we studied a lung trachea segmentation algorithm combining Unet network and attention mechanism, which a convolutional block attention model CBAM focusing on the channel domain and spatial domain was used in the attention mechanism, which improved the tracheal feature weights. In terms of loss function, for the problem of imbalance between positive and negative samples in the original data, this paper used the focal loss function to improve the standard cross-entropy loss function, so that the hard-to-classify samples got more attention in the training process. Finally, the isolated points were removed by eight connected domains judgment, and several larger connected domains were retained, i.e., the last pulmonary trachea part. Twenty-four sets of CT images and 43 sets of CTA images provided by the partner hospitals, totaling 26 157 slice images, were selected as the data set for segmentation experiments. The results showed that the segmentation accuracy reached 0.86, and the mean values of over-segmentation rate and under-segmentation rate were 0.28 and 0.39 respectively. After the ablation experiments of attention module and loss function, the accuracy, over-segmentation rate and under-segmentation rate before improvement were 0.81, 0.30 and 0.40, respectively, indicating the segmentation effect was inferior to the method proposed in this paper. Compared with other commonly used methods under the same conditions, the proposed method reached the highest accuracy rate under the condition that the over-segmentation rate, and under-segmentation rate were guaranteed to be unchanged. The above experiments proved the accuracy of the algorithm in this paper, and successfully solved the problem of inaccurate segmentation of fine trachea.
[1] 谢武英. 权威发布——数据“说”肺癌[J].实用心脑肺血管病杂志, 2019, 27(9): 7-7. [2] 王雷,高欣,张桂芝.形态膨胀的3D区域生长气管分割算法[J].生物医学工程学杂志,2013,30(4):679-683,691. [3] Gao Xin, Zhang Guizhi, Wang Lei.Morphological expansion of 3D region growing trachea segmentation algorithm[J]. Journal of Biomedical Engineering, 2010, 32(2): 470-475. [4] Graham MW, Gibbs JD, Cornish DC, et al.Robust 3-D airway tree segmentation for image-guided peripheral bronchoscopy[J]. IEEE Transactions on Medical Imaging, 2010, 29(4): 982-997. [5] 石跃祥,杜祎.基于T-prim模型的肺气管分割算法[J].计算机应用与软件,2020,37(4):171-177, 185. [6] Xu Ziyue, Bagci U, Foster B, et al.A hybrid multi-scale approach to automatic airway tree segmentation from CT scans[C]//The10th International Symposium on Biomedical Imaging. San Francisco: IEEE, 2013: 1308-1311. [7] 李翠芳, 任彦华, 王远军, 等. 基于CT 图像的肺气管树3D分割方法的研究[J]. 中国医学物理学杂志, 2011, 28(5): 2867-2871. [8] Pu Jiantao, Fuhrman C, Good WF, et al.A differential geometric approach to automated segmentation of human airway tree[J]. IEEE Transactions on Medical Imaging, 2011, 30(2): 266-278. [9] Song Jiangdian, Yang Caiyun, Fan Li, et al.Lung lesion extraction using a toboggan based growing automatic segmentation approach[J]. IEEE Transactions on Medical Imaging, 2015, 35(1): 337-353. [10] 裴宏亮,蒋蕰莙,樊庆文, 等.基于改进区域生长算法提取肺气管树[J].中西医结合心血管病电子杂志,2020,8(21):27-29. [11] Lo P, Sporring J, Ashraf H, et al.Vessel-guided airway tree segmentation: a voxel classification approach[J]. Medical Image Analysis, 2010, 14(4): 527-538. [12] Lee M, Lee JG, Kim N, et al.Hybrid airway segmentation using multi-scale tubular structure filters and texture analysis on 3D chest CT scans[J]. Journal of Digital Imaging, 2019, 32(5): 779-792. [13] Yao Zhenjie, Zhang Zhipeng, Xu Liqun, et al.Convolutional neural network for retinal blood vessel segmentation[C] //International Symposium on Computational Intelligence and Design (ISCID). Hangzhou: IEEE, 2016: 406-409. [14] Wang Shuo, Zhou Mu, Liu Zaiyi, et al.Central focused convolutional neural networks: developing a data-driven model for lung nodule segmentation[J]. Medical Image Analysis, 2017, 40: 172-183. [15] Meng Q, Kitasaka T, Nimura Y, et al.Automatic segmentation of airway tree based on local intensity filter and machine learning technique in 3D chest CT volume[J]. International Journal of Computer Assisted Radiology and Surgery, 2017, 12(2): 245-261. [16] Yun J, Park J, Yu D, et al.Improvement of fully automated airway segmentation on volumetric computed tomographic images using a 2.5 dimensional convolutional neural net[J]. Medical Image Analysis, 2018, 51: 13-21. [17] Juarez AGU, Tiddens HaWM, De Bruijne M. Automatic airway segmentation in chest CT using convolutional neural networks[C] // Lecture Notes in Computer Science. Granada: Springer, 2018: 238-250. [18] 李康. 基于深度学习的CT图像肺气管树分叉点的检测[D].武汉: 华中科技大学,2019. [19] 张琳. 基于自监督迁移学习的肺部气管和血管分割算法研究[D]. 杭州: 浙江大学,2021. [20] 王继伟,王弘轩,黄绍辉,等.基于改进对抗生成网络模型的肺气管图像分割[J].中国数字医学,2021,16(10):93-97. [21] Zhang Guobin, Yang Zhiyong, Huo Bin, et al.Automatic segmentation of organs at risk and tumors in CT images of lung cancer from partially lab elle d datasets with a semi-supervised conditional nnU-Net[J]. Computer Methods and Programs in Biomedicine, 2021, 211:81-90. [22] Tan Wenjun, Huang Peifang, Li Xiaoshuo, et al.Analysis of segmentation of lung parenchyma based on deep learning methods[J]. Journal of X-RAY Science and Technology, 2021, 29(6): 945-959. [23] Malimban J, Lathouwers D, Qian Haibin, et al.Deep learning-based segmentation of the thorax in mouse micro-CT scans[J]. Scientific Reports, 2022, 12(1): 1822. [24] Woo S, Park J, Lee JY, et al.CBAM: convolutional block attention module[C] //Lecture Notes in Computer Science. Munich: Springer-Verlag Berlins, 2018: 3-19. [25] Lin TY, Goyal P, Girshick R, et al.Focal loss for dense object detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(2): 318-327. [26] Ronneberger O, Fischer P, Brox T.U-Net: convolutional networks for biomedical image segmentation[C] // Lecture Notes in Computer Science. Munich: Springer-Verlag Berlin, 2015: 234-241. [27] Zhou Zongwei, Siddiquee MMR, Tajbakhsh N, et al.UNet plus plus: a nested U-Net architecture for medical image segmentation[C] // Lecture Notes in Computer Science. Granada: Springer-Verlag Berlin, 2018: 3-11. [28] Isensee F, Jaeger PF, Kohl SaA, et al. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation[J]. Nature Methods, 2021, 18(2): 203-211. [29] 王瑞绅,宋公飞,王明.引入ECA注意力机制的U-Net语义分割[J].电光与控制,2023,30(1):92-96,102. [30] 兰其斌,黄立勤.基于注意力机制无监督心脏超声序列图像配准[J].福州大学学报(自然科学版),2023,51(1):41-48. [31] 郭鹏,邵剑飞.融合金字塔卷积的U-Net肝脏与肝肿瘤分割网络[J].现代电子技术,2023,46(5):85-88. [32] 徐昌佳. 基于深度学习的结直肠息肉图像分割算法研究[D]. 南昌: 江西理工大学,2022. [33] Kim J, Kang S.Model-agnostic post-processing based on recursive feedback for medical image segmentation[J]. IEEE Access, 2021, 9: 157035-157042. [34] Larrazabal AJ, Martinez C, Ferrante E.Anatomical priors for image segmentation via post-processing with denoising autoencoders[C] //Lecture Notes in Computer Science. Shenzhen: Springer-Verlag Berlin, 2019: 585-593. [35] Isensee F, Jäger PF, Kohl SAA, et al. Automated design of deep learning methods for biomedical image segmentation [EB/OL]. https://arxiv.org/abs/1904.08128, 2019-04-17/2022-01-28.
