Abstract
The shortage of data due to high annotation costs has limited the development of supervised medical image segmentation methods that rely on tight pixel-level annotations. Recently, weakly supervised methods based on multiple instance learning have been proposed to reduce the annotation cost by using bounding box annotations and achieve competitive performance. However, most existing methods require accurate bounding box annotations to generate positive and negative sample bags, which is difficult to realize due to the inevitable errors associated with manually annotated bounding boxes. In this study, we propose a robust framework based on contrast learning for weakly supervised medical image segmentation. Specifically, our method involves a Fine-grained Semantic Representation Module (FSRM), which is used to distinguish foreground and background pixels inside a coarse bounding box. Positive and negative sample bags are generated for multiple instance learning based on the obtained foreground results instead of bounding box constraints. Therefore, our proposed method can ensure the performance under coarse labeling by automatically extracting the boundaries of foreground and background. Our method achieves state-of-the-art results on two publicly available datasets, and extensive experiments validate the robustness of our method under noisy annotations. The source code will be available at https://github.com/ta1ly/wsis-contrastlearning.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chen, J., et al.: TransUNet: transformers make strong encoders for medical image segmentation. arXiv preprint arXiv:2102.04306 (2021)
Hatamizadeh, A., et al.: UNETR: transformers for 3D medical image segmentation. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 574–584 (2022)
Hsu, C.C., Hsu, K.J., Tsai, C.C., Lin, Y.Y., Chuang, Y.Y.: Weakly supervised instance segmentation using the bounding box tightness prior. In: Advances in Neural Information Processing Systems, vol. 32 (2019)
Kervadec, H., Dolz, J., Wang, S., Granger, E., Ayed, I.B.: Bounding boxes for weakly supervised segmentation: global constraints get close to full supervision. In: Medical Imaging with Deep Learning, pp. 365–381. PMLR (2020)
Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)
Krähenbühl, P., Koltun, V.: Efficient inference in fully connected CRFs with gaussian edge potentials. In: Advances in neural information processing systems, vol. 24 (2011)
Kulharia, V., Chandra, S., Agrawal, A., Torr, P., Tyagi, A.: Box2Seg: attention weighted loss and discriminative feature learning for weakly supervised segmentation. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12372, pp. 290–308. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58583-9_18
Li, Z.W., Xuan, S.B., He, X.D., Wang, L.: Global weighted average pooling network with multilevel feature fusion for weakly supervised brain tumor segmentation. IET Image Proc. 17(2), 418–427 (2023)
Liew, S.L., et al.: A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms. Sci. data 9(1), 320 (2022)
Litjens, G., et al.: Evaluation of prostate segmentation algorithms for MRI: the promise12 challenge. Med. Image Anal. 18(2), 359–373 (2014)
Liu, X., et al.: Weakly supervised segmentation of COVID19 infection with scribble annotation on CT images. Pattern Recogn. 122, 108341 (2022)
Mahani, G.K., et al.: Bounding box based weakly supervised deep convolutional neural network for medical image segmentation using an uncertainty guided and spatially constrained loss. In: 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), pp. 1–5. IEEE (2022)
Meng, Q., Liao, L., Satoh, S.: Weakly-supervised learning with complementary heatmap for retinal disease detection. IEEE Trans. Med. Imaging 41(8), 2067–2078 (2022)
Papandreou, G., Chen, L.C., Murphy, K.P., Yuille, A.L.: Weakly-and semi-supervised learning of a deep convolutional network for semantic image segmentation. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1742–1750 (2015)
Paszke, A., et al.: PyTorch: an imperative style, high-performance deep learning library. In: Advances in Neural Information Processing Systems, vol. 32 (2019)
Patel, G., Dolz, J.: Weakly supervised segmentation with cross-modality equivariant constraints. Med. Image Anal. 77, 102374 (2022)
Peng, J., Kervadec, H., Dolz, J., Ayed, I.B., Pedersoli, M., Desrosiers, C.: Discretely-constrained deep network for weakly supervised segmentation. Neural Netw. 130, 297–308 (2020)
Rajchl, M., et al.: DeepCut: object segmentation from bounding box annotations using convolutional neural networks. IEEE Trans. Med. Imaging 36(2), 674–683 (2016)
Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-24574-4_28
Rother, C., Kolmogorov, V., Blake, A.: “GrabCut’’ interactive foreground extraction using iterated graph cuts. ACM Trans. Graph. (TOG) 23(3), 309–314 (2004)
Tajbakhsh, N., Jeyaseelan, L., Li, Q., Chiang, J.N., Wu, Z., Ding, X.: Embracing imperfect datasets: a review of deep learning solutions for medical image segmentation. Med. Image Anal. 63, 101693 (2020)
Wang, J., Xia, B.: Bounding box tightness prior for weakly supervised image segmentation. In: de Bruijne, M., et al. (eds.) MICCAI 2021. LNCS, vol. 12902, pp. 526–536. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-87196-3_49
Wang, J., Xia, B.: Polar transformation based multiple instance learning assisting weakly supervised image segmentation with loose bounding box annotations. arXiv preprint arXiv:2203.06000 (2022)
Wei, J., Hu, Y., Li, G., Cui, S., Kevin Zhou, S., Li, Z.: BoxPolyp: boost generalized polyp segmentation using extra coarse bounding box annotations. In: Wang, L., Dou, Q., Fletcher, P.T., Speidel, S., Li, S. (eds.) Medical Image Computing and Computer Assisted Intervention – MICCAI 2022. MICCAI 2022. Lecture Notes in Computer Science, vol. 13433, pp. 67–77. Springer, Cham (2022). https://doi.org/10.1007/978-3-031-16437-8_7
Xie, J., Xiang, J., Chen, J., Hou, X., Zhao, X., Shen, L.: C2AM: contrastive learning of class-agnostic activation map for weakly supervised object localization and semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 989–998 (2022)
Zhang, Z., Liu, Q., Wang, Y.: Road extraction by deep residual U-Net. IEEE Geosci. Remote Sens. Lett. 15(5), 749–753 (2018)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Zhu, Z., Shi, J., Zhao, M., Wang, Z., Qiao, L., An, H. (2023). Contrast Learning Based Robust Framework for Weakly Supervised Medical Image Segmentation with Coarse Bounding Box Annotations. In: Qin, W., Zaki, N., Zhang, F., Wu, J., Yang, F., Li, C. (eds) Computational Mathematics Modeling in Cancer Analysis. CMMCA 2023. Lecture Notes in Computer Science, vol 14243. Springer, Cham. https://doi.org/10.1007/978-3-031-45087-7_12
Download citation
DOI: https://doi.org/10.1007/978-3-031-45087-7_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-45086-0
Online ISBN: 978-3-031-45087-7
eBook Packages: Computer ScienceComputer Science (R0)
