Skip to main content

Advertisement

SpringerLink
Log in

FlowgateUNet: Dental CT image segmentation network based on FlowFormer and gated attention

  • Original Paper
  • Published:
Signal, Image and Video Processing Aims and scope Submit manuscript

Abstract

Segmentation of cone-beam computed tomography (CBCT) images plays an important role in clinical treatment as well as teaching. Traditional manual segmentation of dental CBCT images requires tools such as mimics and is time-consuming. With the development of deep learning, the U-shaped network, represented by UNet, has shown good results. Due to the significant improvement brought by various applications of Transformers to image tasks, more and more models try to combine attention mechanism with traditional convolutional neural networks. To further improve the performance of dental CBCT segmentation, this paper proposes an improved FlowgateUNet segmentation network, which uses the FlowFormer instead of Transformer in the encoder to achieve attention computation with nearly linear complexity. It also uses the feature map containing global information as the gating signal in the skip connections to further extract relevant features and fuses the results from multiple decoders as the output. Compared to TransUnet, the proposed FlowgateUnet model improved the Dice similarity coefficient (DSC) by 1% on the dental CBCT image dataset, by 0.7% on the dental microCT dataset, and by 2% on the Synapse dataset.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Similar content being viewed by others

Data availibility

The CBCT and microCT datasets are not publicly available due to privacy issues

References

  1. Chen, J., et al.: Transunet: transformers make strong encoders for medical image segmentation. arXiv preprint arXiv:2102.04306 (2021)

  2. Wu, H., Wu, J., Xu, J., Wang, J., Long, M.: Flowformer: linearizing transformers with conservation flows. arXiv preprint arXiv:2202.06258 (2022)

  3. Ronneberger, O., Fischer, P., Brox, T.: U-Net: Convolutional Networks for Biomedical Image Segmentation, pp. 234–241. Springer, Berlin (2015)

    Google Scholar 

  4. Zhou, Z., Rahman Siddiquee, M.M., Tajbakhsh, N., Liang, J.: Unet++: A Nested U-Net Architecture for Medical Image Segmentation, pp. 3–11. Springer, Berlin (2018)

    Google Scholar 

  5. Zhang, Z., Liu, Q., Wang, Y.: Road extraction by deep residual u-net. IEEE Geosci. Remote Sens. Lett. 15, 749–753 (2018)

    Article  ADS  CAS  Google Scholar 

  6. Qin, X., et al.: U2-net: going deeper with nested u-structure for salient object detection. Pattern Recognit. 106, 107404 (2020)

    Article  Google Scholar 

  7. Milletari, F., Navab, N., Ahmadi, S.A.: V-net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation, pp. 565–571. IEEE, Piscataway (2016)

    Google Scholar 

  8. Polizzi, A., et al.: Tooth automatic segmentation from CBCT images: a systematic review. Clin. Oral Investig. 27, 3363–3378 (2023)

    Article  PubMed  Google Scholar 

  9. Cui, Z., Li, C., Wang, W.: Toothnet: automatic tooth instance segmentation and identification from cone beam ct images, 6368–6377 (2019)

  10. Chen, Y., et al.: Automatic segmentation of individual tooth in dental CBCT images from tooth surface map by a multi-task FCN. IEEE Access 8, 97296–97309 (2020)

    Article  Google Scholar 

  11. Guo, M.-H., et al.: Attention mechanisms in computer vision: a survey. Comput. Vis. Media 8, 331–368 (2022)

    Article  Google Scholar 

  12. Wang, X., Girshick, R., Gupta, A., He, K.: Non-local neural networks, pp. 7794–7803 (2018)

  13. Fu, J., et al.: Dual attention network for scene segmentation, pp. 3146–3154 (2019)

  14. Hu, J., Shen, L., Sun, G.: Squeeze-and-excitation networks, pp. 7132–7141 (2018)

  15. Schlemper, J., et al.: Attention gated networks: learning to leverage salient regions in medical images. Med. Image Anal. 53, 197–207 (2019)

    Article  PubMed  PubMed Central  Google Scholar 

  16. Guo, M.-H., et al.: Segnext: rethinking convolutional attention design for semantic segmentation. arXiv preprint arXiv:2209.08575 (2022)

  17. Liu, Z., et al.: A convnet for the 2020s, pp. 11976–11986 (2022)

  18. Vaswani, A., et al.: Attention is all you need. In: Advances in Neural Information Processing Systems, vol. 30 (2017)

  19. Dosovitskiy, A., et al.: An image is worth \(16\times 16\) words: transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

  20. Liu, Z., et al.: Swin transformer: hierarchical vision transformer using shifted windows, pp. 10012–10022 (2021)

  21. Valanarasu, J.M.J., Oza, P., Hacihaliloglu, I., Patel, V.M.: Medical Transformer: Gated Axial-Attention for Medical Image Segmentation, pp. 36–46. Springer, Berlin (2021)

    Google Scholar 

  22. Ji, Y., et al.: Multi-compound Transformer for Accurate Biomedical Image Segmentation, pp. 326–336. Springer, Berlin (2021)

    Google Scholar 

  23. Gao, Y., Zhou, M., Metaxas, D.N.: Utnet: A Hybrid Transformer Architecture for Medical Image Segmentation, pp. 61–71. Springer, Berlin (2021)

    Google Scholar 

  24. Ji, G.P., et al.: Progressively Normalized Self-Attention Network for Video Polyp Segmentation, pp. 142–152. Springer, Berlin (2021)

    Google Scholar 

  25. Zhang, Y., et al.: A Multi-branch Hybrid Transformer Network for Corneal Endothelial Cell Segmentation, pp. 99–108. Springer, Berlin (2021)

    Google Scholar 

  26. Cao, H., et al.: Swin-Unet: Unet-Like Pure Transformer for Medical Image Segmentation, pp. 205–218. Springer, Berlin (2023)

    Google Scholar 

  27. Hatamizadeh, A., et al.: Swin unetr: Swin Transformers for Semantic Segmentation of Brain Tumors in MRI Images, pp. 272–284. Springer, Berlin (2022)

    Google Scholar 

  28. Katharopoulos, A., Vyas, A., Pappas, N., Fleuret, F.: Transformers are RNNs: fast autoregressive transformers with linear attention. In: PMLR, pp. 5156–5165 (2020)

  29. Child, R., Gray, S., Radford, A., Sutskever, I.: Generating long sequences with sparse transformers. arXiv preprint arXiv:1904.10509 (2019)

  30. Wang, S., Li, B. Z., Khabsa, M., Fang, H., Ma, H.: Linformer: self-attention with linear complexity. arXiv preprint arXiv:2006.04768 (2020)

  31. Kitaev, N., Kaiser, Ł., Levskaya, A.: Reformer: the efficient transformer. arXiv preprint arXiv:2001.04451 (2020)

  32. Qin, Z., et al.: cosformer: rethinking softmax in attention. arXiv preprint arXiv:2202.08791 (2022)

  33. Ahujia, R.K., Magnanti, T.L., Orlin, J.B.: Network Flows: Theory, Algorithms and Applications. Prentice-Hall, Hoboken (1993)

    Google Scholar 

  34. Zhou, H.-Y., et al.: nnformer: interleaved transformer for volumetric segmentation. arXiv preprint arXiv:2109.03201 (2021)

  35. Fu, S., et al.: Domain Adaptive Relational Reasoning for 3d Multi-organ Segmentation, pp. 656–666. Springer, Berlin (2020)

    Google Scholar 

  36. Oktay, O., et al.: Attention u-net: learning where to look for the pancreas. arXiv preprint arXiv:1804.03999 (2018)

Download references

Acknowledgements

The research was funded by the National Science Foundation of China under Grant U19A2086 and The Yibin campus major construction and educational reform of CDUT: 22100-000047.

Author information

Authors and Affiliations

  1. College of Computer Science and Cyber Security, Chengdu University of Technology, Chengdu, 610059, China

    Danhua Cao & Biao Cai

  2. College of Industrial Technology, Chengdu University of Technology, Yibin, 644000, China

    Biao Cai

  3. School of Data Science and Artifical Intelligence, Wenzhou University of Technology, Wenzhou, 325035, China

    Mingzhe Liu

Authors
  1. Danhua Cao
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Biao Cai
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Mingzhe Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

DC contributed to conceptualization, methodology, software, writing-original draft. BC contributed to conceptualization, methodology, project administration, supervision, writing-review and editing. ML contributed to investigation, funding acquisition, and resources.

Corresponding author

Correspondence to Biao Cai.

Ethics declarations

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cao, D., Cai, B. & Liu, M. FlowgateUNet: Dental CT image segmentation network based on FlowFormer and gated attention. SIViP 18, 1175–1182 (2024). https://doi.org/10.1007/s11760-023-02765-y

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11760-023-02765-y

Keywords

Search

Navigation