An important prerequisite for improving the reliability of lung cancer surveillance and clinical interventions is accurate lung nodule segmentation. Although deep learning is effective at performing medical image segmentation, lung CT image heterogeneity, nodule size, shape, and location variations, convolutional localized feature extraction characteristics, the receptive field limitations of continuous downsampling, lesion edge information losses, fuzzy boundary segmentation challenges, and the low segmentation accuracy achieved when segmenting lung CT images using deep learning remain. An edge-enhanced multiscale Sobel coordinate attention-atrous spatial convolutional pooling pyramid V-Net (SCA-VNet) algorithm for lung nodule segmentation was proposed to solve these problems. First, a residual edge enhancement module was designed, which was used to enhance the edges of the original data. Using an edge detection operator in combination with a residual module, this module could reduce data redundancy and alleviate the gray level similarity between the foreground and background. Then, a 3D atrous spatial convolutional pooling pyramid module set different expansion rates, which could obtain feature maps under different receptive fields and capture the multiscale information of the segmentation target. Finally, a three-dimensional coordinate attention network (3D CA-Net) module was added to the encoding and decoding paths to extract channel weights from multiple dimensions. This step propagated the spatial information in the coding layer to the subsequent layers, and it could reduce the loss of information during the forward propagation process. The proposed method achieved a Dice coefficient of 87.50% on the lung image database consortium and image database resource initiative (LIDC-IDRI). It significantly outperformed the existing lung nodule segmentation models (UGS-Net, REMU-Net, and multitask models) and compared favorably with the Med3D, CENet, and PCAM_Net segmentation models in terms of their Dice coefficients, which were 3.37%, 2.2%, and 1.43%, respectively. The experimental results showed that the proposed SCA-VNet model attained improved lung nodule segmentation accuracy and laid a good foundation for improving the early detection rate of lung cancer.
Citation: Jinjiang Liu, Yuqin Li, Wentao Li, Zhenshuang Li, Yihua Lan. Multiscale lung nodule segmentation based on 3D coordinate attention and edge enhancement[J]. Electronic Research Archive, 2024, 32(5): 3016-3037. doi: 10.3934/era.2024138
An important prerequisite for improving the reliability of lung cancer surveillance and clinical interventions is accurate lung nodule segmentation. Although deep learning is effective at performing medical image segmentation, lung CT image heterogeneity, nodule size, shape, and location variations, convolutional localized feature extraction characteristics, the receptive field limitations of continuous downsampling, lesion edge information losses, fuzzy boundary segmentation challenges, and the low segmentation accuracy achieved when segmenting lung CT images using deep learning remain. An edge-enhanced multiscale Sobel coordinate attention-atrous spatial convolutional pooling pyramid V-Net (SCA-VNet) algorithm for lung nodule segmentation was proposed to solve these problems. First, a residual edge enhancement module was designed, which was used to enhance the edges of the original data. Using an edge detection operator in combination with a residual module, this module could reduce data redundancy and alleviate the gray level similarity between the foreground and background. Then, a 3D atrous spatial convolutional pooling pyramid module set different expansion rates, which could obtain feature maps under different receptive fields and capture the multiscale information of the segmentation target. Finally, a three-dimensional coordinate attention network (3D CA-Net) module was added to the encoding and decoding paths to extract channel weights from multiple dimensions. This step propagated the spatial information in the coding layer to the subsequent layers, and it could reduce the loss of information during the forward propagation process. The proposed method achieved a Dice coefficient of 87.50% on the lung image database consortium and image database resource initiative (LIDC-IDRI). It significantly outperformed the existing lung nodule segmentation models (UGS-Net, REMU-Net, and multitask models) and compared favorably with the Med3D, CENet, and PCAM_Net segmentation models in terms of their Dice coefficients, which were 3.37%, 2.2%, and 1.43%, respectively. The experimental results showed that the proposed SCA-VNet model attained improved lung nodule segmentation accuracy and laid a good foundation for improving the early detection rate of lung cancer.
| [1] |
Q. Zhou, Y. Fan, Y. Wang, Y. Qiao, G. Wang, Y. Huang, et al., Chinese national guidelines for classification, diagnosis and treatment of pulmonary nodules (2016 version), Chin. J. Lung Cancer, 19 (2016), 793-798. https://doi.org/10.3779/j.issn.1009-3419.2016.12.12 doi: 10.3779/j.issn.1009-3419.2016.12.12
|
| [2] |
T. Dong, L. Wei, S. Nie, Research progress of pulmonary nodule segmentation in CT images, J. Image Graphic, 26 (2021), 751-765. https://doi.org/10.11834/jig.200201 doi: 10.11834/jig.200201
|
| [3] | O. Ronneberger, P. Fischer, T. Brox, U-net: Convolutional networks for biomedical image segmentation, in International Conference on Medical Image Computing and Computer-Assisted Intervention, (2015), 234-241. https://doi.org/10.1007/978-3-319-24574-4_28 |
| [4] | F. Milletari, N. Navab, S. Ahmadi, V-net: Fully convolutional neural networks for volumetric medical image segmentation, in 2016 Fourth International Conference on 3D Vision (3DV), (2016), 565-571. https://doi.org/10.1109/3DV.2016.79 |
| [5] |
D. Li, S. Yuan, G. Yao, Pulmonary nodule segmentation based on REMU-Net, Phys. Eng. Sci. Med., 45 (2022), 995-1004. https://doi.org/10.1007/s13246-022-01157-9 doi: 10.1007/s13246-022-01157-9
|
| [6] |
P. Dutande, U. Baid, S. Talbar, LNCDS: A 2D-3D cascaded CNN approach for lung nodule classification, detection and segmentation, Biomed. Signal Process. Control, 67 (2021), 102527. https://doi.org/10.1016/j.bspc.2021.102527 doi: 10.1016/j.bspc.2021.102527
|
| [7] |
A. A. Sundaresan, A. Jeevanayagam, Efficient multiscale fully convolutional UNet model for segmentation of 3D lung nodule from CT image, J. Med. Imag., 9 (2022), 052402. https://doi.org/10.1117/1.JMI.9.5.052402 doi: 10.1117/1.JMI.9.5.052402
|
| [8] |
Z. Wang, J. Men, F. Zhang, Improved V-Net lung nodule segmentation method based on selective kernel, Signal Image Video Process., 17 (2023), 1763-1774. https://doi.org/10.1007/s11760-022-02387-w doi: 10.1007/s11760-022-02387-w
|
| [9] |
Z. Zhou, F. Gou, Y. Tan, J. Wu, A cascaded multi-stage framework for automatic detection and segmentation of pulmonary nodules in developing countries, IEEE J. Biomed. Health Inf., 26 (2022), 5619-5630. https://doi.org/10.1109/JBHI.2022.3198509 doi: 10.1109/JBHI.2022.3198509
|
| [10] |
W. Zhou, F. Zheng, Y. Zhao, Y. Pang, J. Yi, MSDCNN: A multiscale dilated convolution neural network for fine-grained 3D shape classification, Neural Networks, 172 (2024), 106141. https://doi.org/10.1016/j.neunet.2024.106141 doi: 10.1016/j.neunet.2024.106141
|
| [11] | Y. Wang, W. Zhong, H. Su, F. Zheng, Y. Pang, H. Wen, et al., An improved mvcnn for 3D shape recognition, in 2021 IEEE International Conference on Emergency Science and Information Technology (ICESIT), (2021), 469-472. https://doi.org/10.1109/ICESIT53460.2021.9696941 |
| [12] |
A. Lisowska, Efficient edge detection method for focused images, Appl. Sci., 12 (2022), 11668. https://doi.org/10.3390/app122211668 doi: 10.3390/app122211668
|
| [13] |
A. A. De Borba, A. Muhuri, M. Marengoni, A. C. Frery, Feature selection for edge detection in PolSAR images, Remote Sens., 15 (2023), 2479. https://doi.org/10.3390/rs15092479 doi: 10.3390/rs15092479
|
| [14] |
S. R. Hait, R. Mesiar, P. Gupta, D. Guha, D. Chakraborty, The Bonferroni mean-type pre-aggregation operators construction and generalization: Application to edge detection, Inf. Fusion, 80 (2022), 226-240. https://doi.org/10.1016/j.inffus.2021.11.002 doi: 10.1016/j.inffus.2021.11.002
|
| [15] |
S. Wang, A. Jiang, X. Li, Y. Qiu, M. Li, F. Li, DPBET: A dual-path lung nodules segmentation model based on boundary enhancement and hybrid transformer, Comput. Biol. Med., 151(2022), 106330. https://doi.org/10.1016/j.compbiomed.2022.106330 doi: 10.1016/j.compbiomed.2022.106330
|
| [16] |
Y. Jiang, Pulmonary nodule segmentation algorithm based on Vnet and edge features, Chin. J. Med. Phys., 39 (2022), 705-712. https://doi.org/10.3969/j.issn.1005-202X.2022.06.009 doi: 10.3969/j.issn.1005-202X.2022.06.009
|
| [17] |
Y. Zhang, Y. Li, Y. Zhang, W. Hu, H. Liu, X. Gu, Research progress of laser spot edge detection operator, J. Quantum Opt., (2019), 109-116. https://doi.org/10.3788/JQO20192501.0901 doi: 10.3788/JQO20192501.0901
|
| [18] | Q. Hou, D. Zhou, J. Feng, Coordinate attention for efficient mobile network design, in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (2021), 13713-13722. https://doi.org/10.1109/CVPR46437.2021.01350 |
| [19] |
W. An, G. Wu, Hybrid spatial-channel attention mechanism for cross-age face recognition, Electronics, 13 (2024), 1257. https://doi.org/10.3390/electronics13071257 doi: 10.3390/electronics13071257
|
| [20] |
J. Chen, S. Ye, Z. Jiang, Z. Fang, Image deblurring using feedback mechanism and dual gated attention network, Neural Process. Lett., 56 (2024), 88. https://doi.org/10.1007/s11063-024-11462-x doi: 10.1007/s11063-024-11462-x
|
| [21] |
Y. Li, Z. Zhou, G. Qi, G. Hu, Z. Zhu, X. Huang, Remote sensing micro-object detection under global and local attention mechanism, Remote Sens., 16 (2024), 644. https://doi.org/10.3390/rs16040644 doi: 10.3390/rs16040644
|
| [22] |
Z. Cao, R. Li, X. Yang, L. Fang, Z. Li, J. Li, Multi-scale detection of pulmonary nodules by integrating attention mechanism, Sci. Rep., 13 (2023), 5517. https://doi.org/10.1038/s41598-023-32312-1 doi: 10.1038/s41598-023-32312-1
|
| [23] |
L. C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, A. L. Yuille, Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs, IEEE Trans. Pattern Anal. Mach. Intell., 40 (2017), 834-848. https://doi.org/10.1109/TPAMI.2017.2699184 doi: 10.1109/TPAMI.2017.2699184
|
| [24] |
A. Iii, S. G. Mclennan, G. Bidaut, L. Mcnitt-Gray, M. F. Meyer, C. R. Reeves, et al., The lung image database consortium (LIDC) and image database resource initiative (IDRI): A completed reference database of lung nodules on CT scans, Med. Phys., 38 (2011), 915-931. https://doi.org/10.1118/1.3528204 doi: 10.1118/1.3528204
|
| [25] |
A. A. A. Setio, A. Traverso, T. De Bel, M. S. Berens, C. Van Den Bogaard, P. Cerello, et al., Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: the LUNA16 challenge, Med. Image Anal., 42 (2017), 1-13. https://doi.org/10.1016/j.media.2017.06.015 doi: 10.1016/j.media.2017.06.015
|
| [26] |
Y. Wang, C. Zhou, H. P. Chan, L. M. Hadjiiski, A. Chughtai, E. A. Kazerooni, Hybrid U-Net-based deep learning model for volume segmentation of lung nodules in CT images, Med. Phys., 49 (2022), 7287-7302. https://doi.org/10.1002/mp.15810 doi: 10.1002/mp.15810
|
| [27] | W. Jiang, L. Zhi, S. Zhang, T. Zhou, Segmentation of pulmonary nodules in CT images based on channel residual nested U structure, J. Graphics, 44 (2023), 879-889. http://www.txxb.com.cn/CN/10.11996/JG.j.2095-302X.2023050879 |
| [28] | D. P. Kingma, J. Ba, Adam: A method for stochastic optimization, preprint, arXiv: 1412.6980. |
| [29] | H. Yang, L. Shen, M. Zhang, Q. Wang, Uncertainty-guided lung nodule segmentation with feature-aware attention, in International Conference on Medical Image Computing and Computer-Assisted Intervention, (2022), 44-54. https://doi.org/10.1007/978-3-031-16443-9_5 |
| [30] |
S. Selvadass, P. M. Bruntha, K. M. Sagayam, H. Günerhan, SAtUNet: Series atrous convolution enhanced U-Net for lung nodule segmentation, Int. J. Imaging Syst. Technol., (2023). https://doi.org/10.1002/ima.22964 doi: 10.1002/ima.22964
|
| [31] |
D. Ting, W. Long, Y. Xiaodan, C. Yang, H. Xuewen, N. Shengdong, A full convolution residual network ground-glass pulmonary nodule segmentation method based on empty space convolution pooling pyramid structure and attention mechanism, J. Biomed. Eng., 39 (2022), 11. https://doi.org/10.7507/1001-5515.202010051 doi: 10.7507/1001-5515.202010051
|
| [32] |
X. Yi, X. Jun, X. Gang, X. Xinying, Multi-task pulmonary nodule detection and segmentation with attention feature fusion, Comput. Eng. Design, 43 (2022), 8. https://doi.org/10.16208/j.issn1000-7024.2022.09.017 doi: 10.16208/j.issn1000-7024.2022.09.017
|
| [33] | Z. Li, J. Yang, Y. Xu, L. Zhang, W. Dong, B. Du, Scale-aware test-time click adaptation for pulmonary nodule and mass segmentation, in Medical Image Computing and Computer Assisted Intervention–MICCAI 2023, (2023), 681-691. https://doi.org/10.1007/978-3-031-43898-1_65 |
| [34] |
S. Tyagi, S. N. Talbar, Medicine, CSE-GAN: A 3D conditional generative adversarial network with concurrent squeeze-and-excitation blocks for lung nodule segmentation, Comput. Biol. Med., 147 (2022), 105781. https://doi.org/10.1016/j.compbiomed.2022.105781 doi: 10.1016/j.compbiomed.2022.105781
|
| [35] |
W. Liu, X. Liu, H. Li, M. Li, X. Zhao, Z. Zhu, Integrating lung parenchyma segmentation and nodule detection with deep multi-task learning, IEEE J. Biomed. Health Inf., 25 (2021), 3073-3081. https://doi.org/10.1109/JBHI.2021.3053023 doi: 10.1109/JBHI.2021.3053023
|
| [36] | D. Liang, J. Liu, K. Wang, G. Luo, W. Wang, S. Li, Position-prior clustering-based self-attention module for knee cartilage segmentation, in Medical Image Computing and Computer Assisted Intervention–MICCAI 2022, (2022), 193-202. https://doi.org/10.1007/978-3-031-16443-9_19 |
| [37] | S. Chen, K. Ma, Y. Zheng, Med3D: Transfer learning for 3D medical image analysis, preprint, arXiv: 1904.00625. |
| [38] |
Z. Gu, J. Cheng, H. Fu, K. Zhou, H. Hao, Y. Zhao, et al., Ce-net: Context encoder network for 2d medical image segmentation, IEEE Trans. Med. Imaging, 38 (2019), 2281-2292. https://doi.org/10.1109/TMI.2019.2903562 doi: 10.1109/TMI.2019.2903562
|
Figures(9) / Tables(4)
Jinjiang Liu, Yuqin Li, Wentao Li, Zhenshuang Li, Yihua Lan. Multiscale lung nodule segmentation based on 3D coordinate attention and edge enhancement[J]. Electronic Research Archive, 2024, 32(5): 3016-3037. doi: 10.3934/era.2024138
DownLoad: