Abstract
Automatic segmentation of different organs in fetal Magnetic Resonance Imaging (MRI) plays an important role in measuring the development of the fetus. However, obtaining a large amount of high-quality manually annotated fetal MRI is time-consuming and requires specialized knowledge, which hinders the widespread application that relies on such data to train a model with good segmentation performance. Using weak annotations such as scribbles can substantially reduce the annotation cost, but often leads to poor segmentation performance due to insufficient supervision. In this work, we propose a Scribble-supervised Self-Distillation (ScribSD) method to alleviate this problem. For a student network supervised by scribbles and a teacher based on Exponential Moving Average (EMA), we first introduce prediction-level Knowledge Distillation (KD) that leverages soft predictions of the teacher network to supervise the student, and then propose feature-level KD that encourages the similarity of features between the teacher and student at multiple scales, which efficiently improves the segmentation performance of the student network. Experimental results demonstrate that our KD modules substantially improve the performance of the student, and our method outperforms five state-of-the-art scribble-supervised learning methods.
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
Similar content being viewed by others
References
Budd, S., et al.: Detecting hypo-plastic left heart syndrome in fetal ultrasound via disease-specific atlas maps. In: de Bruijne, M., et al. (eds.) MICCAI 2021. LNCS, vol. 12907, pp. 207–217. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-87234-2_20
Chen, Q., Hong, Y.: Scribble2d5: weakly-supervised volumetric image segmentation via scribble annotations. In: MICCAI, pp. 234–243. Springer Nature Switzerland, Cham (2022). https://doi.org/10.1007/978-3-031-16452-1_23
Fidon, L., et al.: Label-set loss functions for partial supervision: application to fetal brain 3D MRI parcellation. In: de Bruijne, M., et al. (eds.) MICCAI 2021. LNCS, vol. 12902, pp. 647–657. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-87196-3_60
Grandvalet, Y., Bengio, Y.: Semi-supervised learning by entropy minimization. In: NeurIPS, pp. 281–296 (2005)
Hinton, G., Vinyals, O., Dean, J.: Distilling the knowledge in a neural network. In: NeurIPS, pp. 1–9 (2015)
Kim, B., Ye, J.C.: Mumford-Shah loss functional for image segmentation with deep learning. IEEE Trans. Image Process. 29, 1856–1866 (2019)
Lee, H., Jeong, W.-K.: Scribble2Label: scribble-supervised cell segmentation via self-generating pseudo-labels with consistency. In: Martel, A.L., et al. (eds.) MICCAI 2020. LNCS, vol. 12261, pp. 14–23. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-59710-8_2
Lin, D., Dai, J., Jia, J., He, K., Sun, J.: Scribblesup: scribble-supervised convolutional networks for semantic segmentation. In: CVPR, pp. 3159–3167 (2016)
Liu, X., et al.: Weakly supervised segmentation of COVID19 infection with scribble annotation on CT images. Pattern Recogn. 122, 108341 (2022)
Luo, X., et al.: Scribble-supervised medical image segmentation via dual-branch network and dynamically mixed pseudo labels supervision. In: MICCAI, pp. 528–538. Springer (2022). https://doi.org/10.1007/978-3-031-16431-6_50
Luo, X., et al.: WORD: A large scale dataset, benchmark and clinical applicable study for abdominal organ segmentation from CT image. Med. Image Anal. 82, 102642 (2022)
Müller, R., Kornblith, S., Hinton, G.E.: When does label smoothing help? In: NeurIPS, pp. 1–10 (2019)
Obukhov, A., Georgoulis, S., Dai, D., Van Gool, L.: Gated CRF loss for weakly supervised semantic image segmentation. In: NeurIPS, pp. 1–9 (2019)
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
Salehi, S.S.M., et al.: Real-time automatic fetal brain extraction in fetal MRI by deep learning. In: ISBI, pp. 720–724. IEEE (2018)
Tang, M., Djelouah, A., Perazzi, F., Boykov, Y., Schroers, C.: Normalized cut loss for weakly-supervised cnn segmentation. In: CVPR, pp. 1818–1827 (2018)
Tang, M., et al.: On regularized losses for weakly-supervised CNN segmentation. In: Ferrari, V., Hebert, M., Sminchisescu, C., Weiss, Y. (eds.) ECCV 2018. LNCS, vol. 11220, pp. 524–540. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-01270-0_31
Tarvainen, A., Valpola, H.: Mean teachers are better role models: weight-averaged consistency targets improve semi-supervised deep learning results. In NeurIPS, pp. 1195–1204 (2017)
Torrents-Barrena, J., et al.: Fully automatic 3D reconstruction of the placenta and its peripheral vasculature in intrauterine fetal MRI. Med. Image Anal. 54, 263–279 (2019)
Uus, A.U., et al.: Automated 3D reconstruction of the fetal thorax in the standard atlas space from motion-corrupted MRI stacks for 21–36 weeks GA range. Med. Image Anal. 80, 102484 (2022)
Valvano, G., Leo, A., Tsaftaris, S.A.: Learning to segment from scribbles using multi-scale adversarial attention gates. IEEE Trans. Med. Imaging 40(8), 1990–2001 (2021)
Wu, K., Du, B., Luo, M., Wen, H., Shen, Y., Feng, J.: Weakly supervised brain lesion segmentation via attentional representation learning. In: Shen, D., et al. (eds.) MICCAI 2019. LNCS, vol. 11766, pp. 211–219. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-32248-9_24
Xu, K., Rui, L., Li, Y., Gu, L.: Feature normalized knowledge distillation for image classification. In: Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M. (eds.) ECCV 2020. LNCS, vol. 12370, pp. 664–680. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-58595-2_40
Zhang, X., et al.: Confidence-aware cascaded network for fetal brain segmentation on MR images. In: de Bruijne, M., et al. (eds.) MICCAI 2021. LNCS, vol. 12903, pp. 584–593. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-87199-4_55
Acknowledgment
This work was supported by the National Natural Science Foundation of China under Grant 62271115.
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
Qu, Y., Zhao, Q., Wei, L., Lu, T., Zhang, S., Wang, G. (2023). ScribSD: Scribble-Supervised Fetal MRI Segmentation Based on Simultaneous Feature and Prediction Self-distillation. In: Xue, Z., et al. Medical Image Learning with Limited and Noisy Data. MILLanD 2023. Lecture Notes in Computer Science, vol 14307. Springer, Cham. https://doi.org/10.1007/978-3-031-44917-8_2
Download citation
DOI: https://doi.org/10.1007/978-3-031-44917-8_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-47196-4
Online ISBN: 978-3-031-44917-8
eBook Packages: Computer ScienceComputer Science (R0)