In semi-supervised lesion segmentation, the performance of the teacher network is poor, making it difficult for it to guide the student network to perform effective segmentation. To address this issue, an efficient semi-supervised medical image lesion segmentation method was proposed, employing the medical segment anything model (MedSAM), which exhibited superior feature extraction capabilities, as the teacher network. A lightweight student network based on Mamba was constructed, and its segmentation performance was enhanced through knowledge distillation. To address the semantic mismatch caused by feature alignment across heterogeneous networks, a perturbation-consistent cross-architecture knowledge distillation method was introduced. This approach mapped teacher features to the student feature space and aligned perturbation responses, thereby improving the student network’s feature representation ability and improving segmentation performance. Additionally, to tackle the challenges of diverse lesion morphologies and low foreground-background contrast, leading to poor segmentation consistency, a distribution-based self-supervised loss was proposed for optimization. Experiments on multiple types of medical image lesion segmentation datasets demonstrate that the proposed method in this paper outperforms existing methods in segmentation performance. Meanwhile, the student network has only 1.34 M parameters, which significantly improves the model efficiency.
早期方法主要围绕教师-学生框架优化伪标签质量.MT(mean teacher)[1]利用一致性正则化约束学生网络与教师网络的预测一致性,降低标注依赖,但其性能高度依赖教师模型的初始质量,在病灶边界模糊时伪标签噪声显著.UA-MT(uncertainty aware mean teacher)[2]在MT基础上加入不确定性加权损失,缓解噪声敏感性,但未能有效解决病灶多样性带来的偏差.DAMTN(distribution-aware mean teacher network)[3]利用有标数据和无标数据的分布信息来指导模型的学习,以便在训练阶段使模型对有标和无标数据的分割结果的分布尽可能相似,但存在伪标签错误累积循环和双模型计算量大的问题.
TarvainenA, ValpolaH. Mean teachers are better role models: weight-averaged consistency targets improve semi-supervised deep learning results[C]// Proceedings of the International Conference on Learning Representations. Vancouver, 2018:6-11.
[2]
YuL Q, WangS J, LiX M, et al. Uncertainty-aware self-ensembling model for semi-supervised 3D left atrium segmentation[C]//Medical Image Computing and Computer Assisted Intervention—MICCAI 2019. Cham: Springer, 2019: 605-613.
ZhaoXiao-ming, ShiPei-lian, WangDan-dan, et al. Distribution-aware mean teacher networks for semi-supervised medical image segmentation[J]. Journal of Image and Graphics, 2025, 30(2): 575-588.
[5]
LuoX D, LiaoW J, ChenJ N, et al. Efficient semi-supervised gross target volume of nasopharyngeal carcinoma segmentation via uncertainty rectified pyramid consistency[C]//Medical Image Computing and Computer Assisted Intervention—MICCAI 2021. Cham: Springer, 2021: 318-329.
LiFei-xiang, JiangAi-lian. MSMVT: semi-supervised framework with multi-scale and multi-view Transformer for medical image segmentation[J]. Computer Engineering and Applications, 2025, 61(2): 273-282.
[8]
LiS L, ZhangC Y, HeX M. Shape-aware semi-supervised 3D semantic segmentation for medical images[C]//Medical Image Computing and Computer Assisted Intervention—MICCAI 2020. Cham: Springer, 2020: 552-561.
[9]
WuY C, GeZ Y, ZhangD H, et al. Mutual consistency learning for semi-supervised medical image segmentation[J]. Medical Image Analysis, 2022, 81: 102530.
[10]
BaiY H, ChenD W, LiQ L, et al. Bidirectional copy-paste for semi-supervised medical image segmentation[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, 2023: 11514-11524.
[11]
ChenX K, YuanY H, ZengG, et al. Semi-supervised semantic segmentation with cross pseudo supervision[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville, 2021: 2613-2622.
[12]
OualiY, HudelotC, TamiM. Semi-supervised semantic segmentation with cross-consistency training[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle,2020: 12671-12681.
XuXin-hui, ZengZhi-yong, LinZheng-yu. Semi-supervised medical image segmentation method based on consistency regularization[J]. Chinese Journal of Medical Physics, 2025, 42(6): 784-790.
[15]
MiaoJ Z, ChenC, ZhangK L, et al. Cross prompting consistency with segment anything model for semi-supervised medical image segmentation[C]//Medical Image Computing and Computer Assisted Intervention—MICCAI 2024. Cham: Springer, 2024: 167-177.
[16]
ZhangY C, YangJ, LiuY C, et al. SemiSAM: enhancing semi-supervised medical image segmentation via SAM-assisted consistency regularization[C]//2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). Lisbon, 2024: 3982-3986.
[17]
MiaoJ Z, ChenC, LiuF R, et al. CauSSL: causality-inspired semi-supervised learning for medical image segmentation[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris, 2023: 21369-21380.
[18]
WangY C, XiaoB, BiX L, et al. Boundary-aware prototype in semi-supervised medical image segmentation[J]. IEEE Transactions on Image Processing, 2024, 33: 5456-5467.
[19]
MaJ, HeY T, LiF F, et al. Segment anything in medical images[J]. Nature Communications, 2024, 15(1): 654-659.
[20]
LiuZ, LinY T, CaoY, et al. Swin Transformer: hierarchical vision transformer using shifted windows[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Montreal, 2022: 9992-10002.
[21]
HatamizadehA, NathV, TangY C, et al. Swin UNETR: swin transformers for semantic segmentation of brain tumors in MRI images[C]//Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. Cham: Springer, 2022: 272-284.
[22]
GuA, DaoT. Mamba: linear-time sequence modeling with selective state spaces[EB/OL]. (2023-12-01)[2024-12-24].
[23]
GaoJ H, CaiY H, ZhaoZ K, et al. RGU-Mamba: an U-Mamba network with region-based training optimized for domain generalization applied to myocardial scar and edema segmentation[C]//Comprehensive Analysis and Computing of Real-World Medical Images. Cham: Springer, 2025: 77-86.
[24]
JiaoJ B, LiuY, LiuY F, et al. VMamba: visual state space model[C]//Advances in Neural Information Processing Systems. Vancouver, 2024: 103031-103063.
[25]
WuR K, LiuY H, LiangP C, et al. H-vmunet: high-order vision Mamba UNet for medical image segmentation[J]. Neurocomputing, 2025, 624: 129447.
[26]
FanC, YuH Y, HuangY, et al. SliceMamba with neural architecture search for medical image segmentation[J]. IEEE Journal of Biomedical and Health Informatics, 2025, 29(10): 7446-7458.
[27]
HintonG, VinyalsO, DeanJ. Distilling the knowledge in a neural network[EB/OL]. (2015-03-09)[2024-12-24] .
[28]
RomeroA, BallasN, KahouS E, et al. FitNets: hints for thin deep nets[EB/OL]. (2014-12-19)[2024-12-14].
[29]
SunS Q, ChengY, GanZ, et al. Patient knowledge distillation for BERT model compression[EB/OL]. (2019-08-25)[2024-12-24].
[30]
HaoZ, GuoJ, JiaD, et al. Learning efficient vision transformers via fine-grained manifold distillation[C]//Advances in Neural Information Processing Systems. New Orleans, 2022: 9164-9175.
[31]
ZagoruykoS, KomodakisN. Paying more attention to attention: improving the performance of convolutional neural networks via attention transfer[EB/OL]. (2016-12-12)[2024-12-20].
[32]
HaoZ, GuoJ, HanK, et al. One-for-all: bridge the gap between heterogeneous architectures in knowledge distillation[C]//Advances in Neural Information Processing Systems. New Orleans, 2023: 79570-79582.
[33]
LiuY F, CaoJ J, LiB, et al. Cross-architecture knowledge distillation[J]. International Journal of Computer Vision, 2024, 132(8): 2798-2824.
[34]
ZhaiS W, WangG T, LuoX D, et al. PA-seg: learning from point annotations for 3D medical image segmentation using contextual regularization and cross knowledge distillation[J]. IEEE Transactions on Medical Imaging, 2023, 42(8): 2235-2246.