Abstract
Convolutional neural networks (CNN) are widely used in the field of computer-aided diagnosis of lung diseases. Its main tasks are segmentation of lung parenchyma, lung nodule detection and lesion analysis. Among them, the accurate segmentation of lung parenchyma is the key step to achieve lung nodule detection and lung disease diagnosis. Inspired by the self-attention mechanism in transformer, many excellent variants of the attention module combined with convolutional neural networks have been applied to lung image segmentation. However, existing models usually suffer from feature information loss and insensitive multi-scale feature extraction. In this paper, we propose a convolutional neural network ACX-UNet incorporating attention mechanism and cyclic cross-feature extraction in an encoding–decoding mode for lung image segmentation. The convolutional block attention mechanism module (CBAM) is introduced into the skip connection part of the network to reduce the loss of feature information by increasing the attention of high-resolution features; the proposed CX-SPP module is used in the decoding part instead of the two-layer convolution of UNet, and the key information output is cyclically superimposed, deepening the network level and fusing multi-scale information to highlight the target region to suppress the background pixel interference; in order to obtain richer semantic information and improve the generalization ability of the model, the pre-trained VGG16 network is migrated to the encoder part of the UNet model to achieve feature migration and parameter sharing and reduce the training cost. Comparative experiments and ablation experiments are conducted on LIDC and LUNA public datasets with six existing mainstream segmentation networks, and the results show that the prediction maps obtained by the model in this paper are closer to the real labels, and the mIoU, mPA, Precision and Recall evaluation metrics are better than other comparison models.
Similar content being viewed by others
Data availability
The data that support the findings of this study are available on request from the corresponding author.
References
Lv, X., Liang, Wu., Yu, Gu., Zhang, W.L., Jing, L.I.: Detection of low dose CT pulmonary nodules based on 3D convolution neural network. Optics Precis. Eng. 26, 1211–8 (2018). https://doi.org/10.3788/OPE.20182605.1211
Sun, S., Bauer, C., Beichel, R.: Automated 3-D segmentation of lungs with lung cancer in CT data using a novel robust active shape model approach. IEEE Trans. Med. Imaging 31(2), 449–460 (2011)
Liu, C., Zhao, R., Pang, M.: A fully automatic segmentation algorithm for CT lung images based on randomforest. Med. Phys. 47(2), 518–529 (2019). https://doi.org/10.1002/mp.13939
Gao, J., Jiang, Q., Zhou, B., et al.: Convolutional neural networks for computer-aided detection or diagnosis in medical image analysis: an overview. Math. Biosci. Eng. 16(6), 6536–6561 (2019)
Grieser, C., Denecke, T., Rothe, J.H., et al.: Gd-EOB enhanced MRI T1-weighted 3D-GRE with and without elevated flip angle modulation for threshold-based liver volume segmentation. Acta Radiol. 56(12), 1419–1427 (2015)
Zheng, W., Liu, K.: Research on edge detection algorithm in digital image processing. In: 2017 2nd International Conference on Materials Science, Machinery and Energy Engineering (MSMEE 2017). Atlantis Press, 2017: 1203–1208.
Anshad, P.Y.M., Kumar, S.S., Shahudheen, S.: Segmentation of chondroblastoma from medical images using modified region growing algorithm. Clust. Comput. 22, 13437–13444 (2019)
Long, J., Shelhamer, E., Darrell, T.: Fully convolutional networks for semantic segmentation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 3431–3440. (2015)
Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedical image segmentation. In: Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18. Springer International Publishing, pp. 234–241 (2015)
Skourt, B.A., El Hassani, A., Majda, A.: Lung CT image segmentation using deep neural networks. Proc. Comput. Sci. 127, 109–113 (2018)
Jin, Q., Meng, Z., Sun, C., et al.: RA-UNet: a hybrid deep attention-aware network to extract liver and tumor in CT scans. Front. Bioeng. Biotechnol. 8, 1471 (2020)
Xiao, Z., Liu, B., Geng, L., et al.: Segmentation of lung nodules using improved 3D-UNet neural network. Symmetry 12(11), 1787 (2020)
Zezhi, W., Li, X., Zuo, J.: RAD-UNet: research on an improved lung nodule semantic segmentation algorithm based on deep learning. Front. Oncol. (2023). https://doi.org/10.3389/fonc.2023.1084096
Jiang, F., Gu, Q., Hao, H.Z., et al.: Survey on content-based image segmentation methods. J. Softw. 28(1), 160–183 (2016)
Liu, J., Chen, A., Zhou, G., et al.: Dermatoscopic image melanoma recognition based on CFLDnet fusion network. Multimed. Tools Appl. 80, 25477–25494 (2021)
Chaudhari, S., Mithal, V., Polatkan, G., et al.: An attentive survey of attention models. ACM Trans. Intell. Syst. Technol. 12(5), 1–32 (2021)
Tsotsos, J.K., Culhane, S.M., Wai, W.Y.K., et al.: Modeling visual attention via selective tuning. Artif. Intell. 78(1–2), 507–545 (1995)
Jiang, D., Sun, B., Su, S., et al.: FASSD: a feature fusion and spatial attention-based single shot detector for small object detection. Electronics 9(9), 1536 (2020)
Fu, J., Liu, J., Tian, H., et al.: Dual attention network for scene segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 3146–3154. (2019)
Li, X., Jiang, Y., Li, M., et al.: Lightweight attention convolutional neural network for retinal vessel image segmentation. IEEE Trans. Industr. Inf. 17(3), 1958–1967 (2020)
Chen, L.C., Papandreou, G., Kokkinos, I., et al.: Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE Trans. Pattern Anal. Mach. Intell. 40(4), 834–848 (2017)
Vaswani, A. et al.: Attention is all you need. In: 2017 Proc. Adv. Neural Inf.Process. Syst., vol. 30, pp. 1–11. (2017)
Mishra, S., Zhang, Y., Chen, D.Z., et al.: Data-driven deep supervision for medical image segmentation. IEEE Trans. Med. Imaging 41(6), 1560–1574 (2022)
Milletari, F., Navab, N., Ahmadi, S. A.: V-net: Fully convolutional neural networks for volumetric medical image segmentation. In: 2016 Fourth International Conference on 3D Vision (3DV). IEEE, pp. 565–571. (2016)
Kingma, D. P., Ba, J.: Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, (2014).
Trebing, K., Staǹczyk, T., Mehrkanoon, S.: SmaAt-UNet: precipitation nowcasting using a small attention-UNet architecture. Pattern Recogn. Lett. 145, 178–186 (2021)
Chen, X., Zhang, R., Yan, P.: Feature fusion encoder decoder network for automatic liver lesion segmentation. In: 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019). IEEE, 2019: 430-433
Li, H., Xiong, P., Fan, H., et al.: Dfanet: Deep feature aggregation for real-time semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2019: 9522–9531.
Funding
National Natural Science Foundation of China (61972456,61173032); Tianjin Research Innovation Project for Postgraduate Students (2022SKY126); National Student Innovation Training Program(202310058029).
Author information
Authors and Affiliations
Contributions
HW, ZZ, and YZ wrote the main manuscript text; BS and XZ made critical revisions to the paper and provided fund support. All authors reviewed the manuscript.
Corresponding authors
Ethics declarations
Conflict of interest
The authors declare no competing interest.
Ethical approval
This article does not involve ethical issues.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Wu, H., Zhang, Z., Zhang, Y. et al. ACX-UNet: a multi-scale lung parenchyma segmentation study with improved fusion of skip connection and circular cross-features extraction. SIViP 18, 525–533 (2024). https://doi.org/10.1007/s11760-023-02770-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11760-023-02770-1