Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

FPL+: Filtered Pseudo Label-based Unsupervised Cross-Modality Adaptation for 3D Medical Image Segmentation (2404.04971v1)

Published 7 Apr 2024 in cs.CV

Abstract: Adapting a medical image segmentation model to a new domain is important for improving its cross-domain transferability, and due to the expensive annotation process, Unsupervised Domain Adaptation (UDA) is appealing where only unlabeled images are needed for the adaptation. Existing UDA methods are mainly based on image or feature alignment with adversarial training for regularization, and they are limited by insufficient supervision in the target domain. In this paper, we propose an enhanced Filtered Pseudo Label (FPL+)-based UDA method for 3D medical image segmentation. It first uses cross-domain data augmentation to translate labeled images in the source domain to a dual-domain training set consisting of a pseudo source-domain set and a pseudo target-domain set. To leverage the dual-domain augmented images to train a pseudo label generator, domain-specific batch normalization layers are used to deal with the domain shift while learning the domain-invariant structure features, generating high-quality pseudo labels for target-domain images. We then combine labeled source-domain images and target-domain images with pseudo labels to train a final segmentor, where image-level weighting based on uncertainty estimation and pixel-level weighting based on dual-domain consensus are proposed to mitigate the adverse effect of noisy pseudo labels. Experiments on three public multi-modal datasets for Vestibular Schwannoma, brain tumor and whole heart segmentation show that our method surpassed ten state-of-the-art UDA methods, and it even achieved better results than fully supervised learning in the target domain in some cases.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (47)
  1. H. R. Roth, C. Shen, H. Oda, M. Oda et al., “Deep learning and its application to medical image segmentation,” Medical Imaging Technology, vol. 36, no. 2, pp. 63–71, 2018.
  2. U. Baid, S. Ghodasara, S. Mohan, M. Bilello et al., “The RSNA-ASNR-MICCAI BraTS 2021 benchmark on brain tumor segmentation and radiogenomic classification,” arXiv preprint arXiv:2107.02314, 2021.
  3. G. Wang, J. Shapey, W. Li et al., “Automatic segmentation of vestibular schwannoma from T2-weighted MRI by deep spatial attention with hardness-weighted loss,” in MICCAI, 2019, pp. 264–272.
  4. G. Wang, W. Li, M. A. Zuluaga et al., “Interactive medical image segmentation using deep learning with image-specific fine tuning,” IEEE transactions on medical imaging, vol. 37, no. 7, pp. 1562–1573, 2018.
  5. J. Donahue, J. Hoffman, E. Rodner et al., “Semi-supervised domain adaptation with instance constraints,” in CVPR, 2013, pp. 668–675.
  6. R. Gu, J. Zhang, G. Wang, W. Lei, T. Song, X. Zhang, K. Li, and S. Zhang, “Contrastive semi-supervised learning for domain adaptive segmentation across similar anatomical structures,” IEEE Transactions on Medical Imaging, vol. 42, no. 1, pp. 245–256, 2022.
  7. J.-Y. Zhu, T. Park, P. Isola, and A. A. Efros, “Unpaired image-to-image translation using cycle-consistent adversarial networks,” in ICCV, 2017, pp. 2223–2232.
  8. T. Park, A. A. Efros, R. Zhang, and J.-Y. Zhu, “Contrastive learning for unpaired image-to-image translation,” in ECCV, 2020, pp. 319–345.
  9. J. Wu, D. Guo, L. Wang, S. Yang, Y. Zheng, J. Shapey, T. Vercauteren, S. Bisdas, R. Bradford, S. Saeed et al., “TISS-Net: Brain tumor image synthesis and segmentation using cascaded dual-task networks and error-prediction consistency,” Neurocomputing, p. 126295, 2023.
  10. E. Tzeng, J. Hoffman, N. Zhang, K. Saenko, and T. Darrell, “Deep domain confusion: Maximizing for domain invariance,” arXiv preprint arXiv:1412.3474, 2014.
  11. M. Long, H. Zhu, J. Wang, and M. I. Jordan, “Unsupervised domain adaptation with residual transfer networks,” in NeurIPS, 2016, p. 136–144.
  12. J. Hoffman, E. Tzeng, T. Park, J.-Y. Zhu, P. Isola, K. Saenko, A. Efros, and T. Darrell, “Cycada: Cycle-consistent adversarial domain adaptation,” in ICML, 2018, pp. 1989–1998.
  13. C. Chen, Q. Dou, H. Chen, J. Qin, and P. A. Heng, “Unsupervised bidirectional cross-modality adaptation via deeply synergistic image and feature alignment for medical image segmentation,” IEEE transactions on medical imaging, vol. 39, no. 7, pp. 2494–2505, 2020.
  14. X. Han, L. Qi, Q. Yu, Z. Zhou, Y. Zheng, Y. Shi, and Y. Gao, “Deep symmetric adaptation network for cross-modality medical image segmentation,” IEEE transactions on medical imaging, vol. 41, no. 1, pp. 121–132, 2021.
  15. T.-H. Vu, H. Jain, M. Bucher, M. Cord, and P. Pérez, “Advent: Adversarial entropy minimization for domain adaptation in semantic segmentation,” in CVPR, 2019, pp. 2517–2526.
  16. X. Luo, M. Hu, W. Liao, S. Zhai, T. Song, G. Wang, and S. Zhang, “Scribble-supervised medical image segmentation via dual-branch network and dynamically mixed pseudo labels supervision,” in MICCAI, 2022, pp. 528–538.
  17. Y. Li, J. Chen, X. Xie, K. Ma, and Y. Zheng, “Self-loop uncertainty: A novel pseudo-label for semi-supervised medical image segmentation,” in MICCAI, 2020, pp. 614–623.
  18. Y. Lin, H. Yao, Z. Li, G. Zheng, and X. Li, “Calibrating label distribution for class-imbalanced barely-supervised knee segmentation,” in MICCAI, 2022, pp. 109–118.
  19. G. Wang, X. Luo, R. Gu, S. Yang, Y. Qu, S. Zhai, Q. Zhao, K. Li, and S. Zhang, “Pymic: A deep learning toolkit for annotation-efficient medical image segmentation,” Comput Methods Programs Biomed, vol. 231, p. 107398, 2023.
  20. J. Wu, R. Gu, G. Dong, G. Wang, and S. Zhang, “FPL-UDA: Filtered pseudo label-based unsupervised cross-modality adaptation for vestibular schwannoma segmentation,” in ISBI, 2022, pp. 1–5.
  21. Q. Dou, C. Ouyang, C. Chen, H. Chen, B. Glocker, X. Zhuang, and P.-A. Heng, “Pnp-adanet: Plug-and-play adversarial domain adaptation network at unpaired cross-modality cardiac segmentation,” IEEE Access, vol. 7, pp. 99 065–99 076, 2019.
  22. F. Wu and X. Zhuang, “CF distance: a new domain discrepancy metric and application to explicit domain adaptation for cross-modality cardiac image segmentation,” IEEE Transactions on Medical Imaging, vol. 39, no. 12, pp. 4274–4285, 2020.
  23. Y. Yao, F. Liu, Z. Zhou, Y. Wang, W. Shen, A. Yuille, and Y. Lu, “Unsupervised domain adaptation through shape modeling for medical image segmentation,” in MIDL, 2022, pp. 1444–1458.
  24. K. Yao, Z. Su, K. Huang, X. Yang, J. Sun, A. Hussain, and F. Coenen, “A novel 3D unsupervised domain adaptation framework for cross-modality medical image segmentation,” IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 10, pp. 4976–4986, 2022.
  25. Y. Wu, Z. Ge, D. Zhang, M. Xu, L. Zhang, Y. Xia, and J. Cai, “Mutual consistency learning for semi-supervised medical image segmentation,” Medical Image Analysis, vol. 81, p. 102530, 2022.
  26. A. Ghosh, H. Kumar, and P. S. Sastry, “Robust loss functions under label noise for deep neural networks,” in AAAI, 2017, p. 1919–1925.
  27. Z. Zhang and M. Sabuncu, “Generalized cross entropy loss for training deep neural networks with noisy labels,” in NeurIPS, 2018, p. 8778–8788.
  28. G. Wang, X. Liu, C. Li, Z. Xu, J. Ruan, H. Zhu, T. Meng, K. Li, N. Huang, and S. Zhang, “A noise-robust framework for automatic segmentation of COVID-19 pneumonia lesions from CT images,” IEEE Transactions on Medical Imaging, vol. 39, no. 8, pp. 2653–2663, 2020.
  29. Z. Xu, D. Lu, J. Luo, Y. Wang, J. Yan, K. Ma, Y. Zheng, and R. K.-Y. Tong, “Anti-interference from noisy labels: Mean-teacher-assisted confident learning for medical image segmentation,” IEEE Transactions on Medical Imaging, vol. 41, no. 11, pp. 3062–3073, 2022.
  30. G. Wang, S. Zhai, G. Lasio, B. Zhang, B. Yi, S. Chen et al., “Semi-supervised segmentation of radiation-induced pulmonary fibrosis from lung CT scans with multi-scale guided dense attention,” IEEE Transactions on Medical Imaging, vol. 41, no. 3, pp. 531–542, 2022.
  31. B. Han, Q. Yao, X. Yu, G. Niu, M. Xu, W. Hu, I. Tsang, and M. Sugiyama, “Co-teaching: Robust training of deep neural networks with extremely noisy labels,” in NeurIPS, 2018, p. 8527–8537.
  32. T. Zhang, L. Yu et al., “Robust medical image segmentation from non-expert annotations with tri-network,” in MICCAI, 2020, pp. 249–258.
  33. S. Yang, G. Wang, H. Sun, X. Luo et al., “Learning covid-19 pneumonia lesion segmentation from imperfect annotations via divergence-aware selective training,” IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 8, pp. 3673–3684, 2022.
  34. R. Gu, J. Zhang, G. Wang, W. Lei, T. Song, X. Zhang, K. Li, and S. Zhang, “Contrastive semi-supervised learning for domain adaptive segmentation across similar anatomical structures,” IEEE Transactions on Medical Imaging, vol. 42, no. 1, pp. 245–256, 2023.
  35. Z. Zhou, L. Qi, X. Yang, D. Ni, and Y. Shi, “Generalizable cross-modality medical image segmentation via style augmentation and dual normalization,” in CVPR, 2022, pp. 20 856–20 865.
  36. W.-G. Chang, T. You, S. Seo, S. Kwak, and B. Han, “Domain-specific batch normalization for unsupervised domain adaptation,” in CVPR, 2019, pp. 7354–7362.
  37. S. Ioffe and C. Szegedy, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” in ICML, 2015, pp. 448–456.
  38. Y. Gal and Z. Ghahramani, “Dropout as a bayesian approximation: Representing model uncertainty in deep learning,” in ICML, 2016, p. 1050–1059.
  39. J. Shapey, A. Kujawa, R. Dorent, G. Wang et al., “Segmentation of vestibular schwannoma from MRI, an open annotated dataset and baseline algorithm,” Scientific Data, vol. 8, no. 1, pp. 1–6, 2021.
  40. R. Dorent, A. Kujawa, M. Ivory, S. Bakas et al., “Crossmoda 2021 challenge: Benchmark of cross-modality domain adaptation techniques for vestibular schwannoma and cochlea segmentation,” Medical Image Analysis, vol. 83, p. 102628, 2023.
  41. B. Zhou, C. Liu, and J. S. Duncan, “Anatomy-constrained contrastive learning for synthetic segmentation without ground-truth,” in MICCAI, 2021, pp. 47–56.
  42. L. Hoyer, D. Dai, and L. Van Gool, “HRDA: Context-aware high-resolution domain-adaptive semantic segmentation,” in ECCV, 2022, pp. 372–391.
  43. K. Wang, D. Kim, R. Feris, and M. Betke, “CDAC: Cross-domain attention consistency in transformer for domain adaptive semantic segmentation,” in ICCV, 2023, pp. 11 519–11 529.
  44. L. Hoyer, D. Dai, H. Wang, and L. Van Gool, “MIC: Masked image consistency for context-enhanced domain adaptation,” in CVPR, 2023, pp. 11 721–11 732.
  45. B. H. Menze, A. Jakab, S. Bauer et al., “The multimodal brain tumor image segmentation benchmark (BRATS),” IEEE transactions on medical imaging, vol. 34, no. 10, pp. 1993–2024, 2014.
  46. X. Luo and X. Zhuang, “𝒳𝒳\mathcal{X}caligraphic_X-metric: An N-dimensional information-theoretic framework for groupwise registration and deep combined computing,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 7, pp. 9206–9224, 2023.
  47. J. Xian, X. Li et al., “Unsupervised cross-modality adaptation via dual structural-oriented guidance for 3D medical image segmentation,” IEEE Transactions on Medical Imaging, vol. 42, no. 6, pp. 1774–1785, 2023.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Jianghao Wu (8 papers)
  2. Dong Guo (46 papers)
  3. Guotai Wang (67 papers)
  4. Qiang Yue (5 papers)
  5. Huijun Yu (2 papers)
  6. Kang Li (207 papers)
  7. Shaoting Zhang (133 papers)
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now