基于可疑像素相互修正的半监督医学图像分割

杨子瑶; 雷涛; 杜晓刚; 王梦溪; 孙菲蔓; 公茂果

doi:10.12263/DZXB.20240616

您当前的位置：

首页 >

文章列表页 >

基于可疑像素相互修正的半监督医学图像分割

学术论文 | 更新时间：2025-08-18

- 基于可疑像素相互修正的半监督医学图像分割
  增强出版
- Semi-Supervised Medical Image Segmentation Based on Suspicious Pixel Mutual Correction
- 电子学报 2025年53卷第5期页码：1607-1621
- 作者机构：
  
  1.陕西科技大学电子信息与人工智能学院，陕西西安 710021
  2.陕西省人工智能联合实验室（陕西科技大学），陕西西安 710021
  3.西南交通大学数学学院，四川成都 611756
  4.西安电子科技大学协同智能系统教育部重点实验室，陕西西安 710071
- 作者简介：
  
  [ "杨子瑶女，2002年9月出生于陕西省西安市.现为陕西科技大学电子信息与人工智能学院硕士研究生.主要研究方向为计算机视觉.E-mail: 231611038@sust.edu.cn" ]
  [ "雷涛男，1981年11月出生于陕西省渭南市.2011年在西北工业大学获得博士学位，现为陕西科技大学教授，博士生导师.主要研究方向为图像处理、模式识别和计算机视觉.中国电子学会会员编号：E190184479M.E-mail: leitao@sust.edu.cn" ]
  [ "杜晓刚男，1985年9月出生于陕西省宝鸡市.现为陕西科技大学电子信息与人工智能学院副教授.主要研究方向为机器学习、计算机视觉、医学图像处理.E-mail: duxiaogang@sust.edu.cn" ]
- 基金信息：
  
  国家自然科学基金(62271296;62201334);陕西省教育厅科学研究计划项目(23JP022;23JP014)
- DOI：10.12263/DZXB.20240616
  中图分类号： TP391;
- 收稿：2024-07-02，
  
  修回：2025-04-01，
  
  纸质出版：2025-05-25
- 稿件说明：
移动端阅览
杨子瑶, 雷涛, 杜晓刚, 等. 基于可疑像素相互修正的半监督医学图像分割[J]. 电子学报, 2025, 53(05): 1607-1621.

YANG Zi-yao, LEI Tao, DU Xiao-gang, et al. Semi-Supervised Medical Image Segmentation Based on Suspicious Pixel Mutual Correction[J]. Acta Electronica Sinica, 2025, 53(05): 1607-1621.
杨子瑶, 雷涛, 杜晓刚, 等. 基于可疑像素相互修正的半监督医学图像分割[J]. 电子学报, 2025, 53(05): 1607-1621. DOI：10.12263/DZXB.20240616

YANG Zi-yao, LEI Tao, DU Xiao-gang, et al. Semi-Supervised Medical Image Segmentation Based on Suspicious Pixel Mutual Correction[J]. Acta Electronica Sinica, 2025, 53(05): 1607-1621. DOI：10.12263/DZXB.20240616

摘要

现有的半监督学习方法通常对图像中的所有像素进行同等处理，忽视了图像内不同区域的复杂度差异.这导致模型对于预测难度较高的挑战性区域学习不足，从而降低了其对这些区域的处理能力.此外，由于伪标签是基于当前模型对未标注数据的预测结果生成的，而模型在挑战性区域表现较差，因此不准确的预测增加了伪标签中引入噪声的风险，进而降低了伪标签的可信度.针对上述问题，提出了一种基于可疑像素相互修正的半监督学习框架（Suspicious Pixel Mutual Correction，SPMC）.该框架由两个编码器结构相同，但解码器上采样方式不同的网络构成.首先，设计了一个用于处理标注数据的共困像素筛查（Common Difficulty Pixel Screening，CDPS）模块.通过利用每个网络预测结果中的可疑像素，精准筛查出两个网络都预测困难的共困像素，并使用准确的监督信号对其进行修正，从而提高模型对挑战性区域预测的准确性.其次，设计了一个用于处理未标注数据的软伪标签辅助教学（Soft Pseudo-Label Assisted Teaching，SPLAT）模块.通过利用一个网络生成的软伪标签中的可信像素，选择性地对另一个网络预测结果中的可疑像素进行伪监督.两个网络通过高质量交互来更新参数，从而减少模型的认知偏差并提升伪标签质量.实验结果表明，提出的方法在三种公开医学数据集左心房（Left Atrium，LA）、脑部肿瘤分割（Brain Tumor Segmentation，BraTS）和自动心脏挑战（Automatic Cardiac Diagnosis Challenge，ACDC）上的性能均优于当前主流的半监督学习方法.

Abstract

Existing semi-supervised learning methods typically handle all pixels in the image equally

ignoring the differences in complexity of different regions within the image. This results in the model’s insufficient learning of challenging regions with higher difficulty to predict

reducing its ability to process challenging areas. Furthermore

since pseudo-labels are generated based on the model’s predictions on unlabeled data

and the model performs poorly in challenging regions

inaccurate predictions increase the risk of introducing noise into the pseudo-labels

thereby reducing their reliability. To address these issues

a semi-supervised learning framework based on suspicious pixels mutual correction (SPMC) is proposed

the framework consists of two networks with identical encoder structures but different upsampling methods in the decoder. Firstly

a common difficulty pixel screening (CDPS) module was designed to handle labeled data. By utilizing suspicious pixels from the prediction results of eatch network to accurately screen out the hard to predict pixels in both networks. These pixels are then corrected using precise supervision signals

thereby improving the model’s prediction accuracy in challenging regions. Secondly

a soft pseudo-label assisted teaching (SPLAT) module was developed to handle unlabeled data. By utilizing trusted pixels from the soft pseudo-labels generated by one network to selectively pseudo-supervise suspicious pixels in the predicted results of the other network. Two networks update parameters through high-quality interaction

thereby reducing the model's cognitive bias and enhancing the quality of the pseudo-labels. Experimental results on three publicly available medical datasets

left atrium (LA)

brain tumor segmentation (BraTS)

and automatic cardiac diagnosis challenge (ACDC)

show that the proposed method is superior to current mainstream semi-supervised learning methods.

关键词

Keywords

references

WANG R S , LEI T , CUI R X , et al . Medical image segmentation using deep learning: A survey [J ] . IET Image Processing , 2022 , 16 ( 5 ): 1243 - 1267 .

RONNEBERGER O , FISCHER P , BROX T . U-Net: Convolutional networks for biomedical image segmentation [M ] // Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015 . Cham : Springer International Publishing , 2015 : 234 - 241 .

ÇIÇEK Ö , ABDULKADIR A , LIENKAMP S S , et al . 3D U-Net: Learning dense volumetric segmentation from sparse annotation [M ] // Medical Image Computing and Computer-Assisted Intervention-MICCAI 2016 . Cham : Springer International Publishing , 2016 : 424 - 432 .

ZHOU Z W , SIDDIQUEE M M R , TAJBAKHSH N , et al . UNet++: Redesigning skip connections to exploit multiscale features in image segmentation [J ] . IEEE Transactions on Medical Imaging , 2020 , 39 ( 6 ): 1856 - 1867 .

LI X Q , HE Q , DAI S M , et al . Semi-supervised semantic segmentation constrained by consistency regularization [C ] // 2020 IEEE International Conference on Multimedia and Expo (ICME) . Piscataway : IEEE , 2020 : 1 - 6 .

LI X M , YU L Q , CHEN H , et al . Transformation-consistent self-ensembling model for semisupervised medical image segmentation [J ] . IEEE Transactions on Neural Networks and Learning Systems , 2021 , 32 ( 2 ): 523 - 534 .

SHEN Z Q , CAO P , YANG H , et al . Co-training with high-confidence pseudo labels for semi-supervised medical image segmentation [C ] // Proceedings of the 32nd International Joint Conference on Artificial Intelligence . Barcelona : International Joint Conferences on Artificial Intelligence Organization , 2023 : 4199 - 4207 .

LU S F , YAN Z Y , CHEN W , et al . Dual consistency regularization with subjective logic for semi-supervised medical image segmentation [J ] . Computers in Biology and Medicine , 2024 , 170 : 107991 .

CHEN Y T , MA Y , MEI X G , et al . Triple-task mutual consistency for semi-supervised 3D medical image segmentation [J ] . Computers in Biology and Medicine , 2024 , 175 : 108506 .

WU Y C , XU M F , GE Z Y , et al . Semi-supervised left atrium segmentation with mutual consistency training [M ] // Medical Image Computing and Computer Assisted Intervention-MICCAI 2021 . Cham : Springer International Publishing , 2021 : 297 - 306 .

TARVAINEN A , VALPOLA H . Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results [J ] . Advances in Neural Information Processing Systems , 2017 , 30 : 1 - 10 .

LIU Y Y , TIAN Y , CHEN Y H , et al . Perturbed and strict mean teachers for semi-supervised semantic segmentation [C ] // 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2022 : 4248 - 4257 .

SOHN K , BERTHELOT D , LI C L , et al . FixMatch: Simplifying semi-supervised learning with consistency and confidence [EB/OL ] . ( 2020-11-25 )[ 2024-07-02 ] . https://arxiv.org/abs/2001.07685v2 https://arxiv.org/abs/2001.07685v2 .

YANG L H , QI L , FENG L T , et al . Revisiting weak-to-strong consistency in semi-supervised semantic segmentation [C ] // 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2023 : 7236 - 7246 .

JIN G , LIU C C , CHEN X . Adversarial network integrating dual attention and sparse representation for semi-supervised semantic segmentation [J ] . Information Processing & Management , 2021 , 58 ( 5 ): 102680 .

XU D , WANG Z L . Semi-supervised semantic segmentation using an improved generative adversarial network [J ] . Journal of Intelligent & Fuzzy Systems , 2021 , 40 ( 5 ): 9709 - 9719 .

黄子臻 , 王雷 , 张玉坤 , 等 . 用于医学图像分割的半监督对抗自集成网络 [J/OL ] . 计算机辅助设计与图形学学报 . ( 2024-10-29 )[ 2025-04-01 ] . https://link.cnki.net/urlid/11.2925.TP.20241029.1352.024 https://link.cnki.net/urlid/11.2925.TP.20241029.1352.024 .

HUANG Z Z , WANG L , ZHANG Y K , et al . Semi-supervised adversarial self-integration network for medical image segmentation [J/OL ] . Journal of Computer-Aided Design & Computer Graphics . ( 2024-10-29 )[ 2025-04-01 ] . https://link.cnki.net/urlid/11.2925.TP.20241029.1352.024. https://link.cnki.net/urlid/11.2925.TP.20241029.1352.024. (in Chinese)

WANG W G , ZHOU T F , YU F , et al . Exploring cross-image pixel contrast for semantic segmentation [C ] // 2021 IEEE/CVF International Conference on Computer Vision (ICCV) . Piscataway : IEEE , 2021 : 7283 - 7293 .

ZHONG Y Y , YUAN B D , WU H , et al . Pixel contrastive-consistent semi-supervised semantic segmentation [C ] // 2021 IEEE/CVF International Conference on Computer Vision (ICCV) . Piscataway : IEEE , 2021 : 7253 - 7262 .

TANG C , ZENG X Y , ZHOU L P , et al . Semi-supervised medical image segmentation via hard positives oriented contrastive learning [J ] . Pattern Recognition , 2024 , 146 : 110020 .

LEI T , LIU H L , CHEN Q , et al . Semi-supervised 3D medical image segmentation using shape-guided dual consistency learning [C ] // 2022 IEEE International Conference on Multimedia and Expo (ICME) . Piscataway : IEEE , 2022 : 1 - 6 .

LEE D H . Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks [C ] // Workshop on Challenges in Representation Learning . Atlanta : ICML . 2013 : 896 .

YANG L H , ZHUO W , QI L , et al . ST: Make self-training work better for semi-supervised semantic segmentation [C ] // 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2022 : 4258 - 4267 .

魏翔 , 王靖杰 , 张顺利 , 等 . ReLSL: 基于可靠标签选择与学习的半监督学习算法 [J ] . 计算机学报 , 2022 , 45 ( 6 ): 1147 - 1160 .

WEI X , WANG J J , ZHANG S L , et al . ReLSL: Reliable label selection and learning based algorithm for semi-supervised learning [J ] . Chinese Journal of Computers , 2022 , 45 ( 6 ): 1147 - 1160 . (in Chinese)

LEI T , ZHANG D , DU X G , et al . Semi-supervised medical image segmentation using adversarial consistency learning and dynamic convolution network [J ] . IEEE Transactions on Medical Imaging , 2023 , 42 ( 5 ): 1265 - 1277 .

XIAO H , LI D , XU H , et al . Semi-supervised semantic segmentation with cross teacher training [J ] . Neurocomputing , 2022 , 508 : 36 - 46 .

方超伟 , 李雪 , 李钟毓 , 等 . 基于双模型交互学习的半监督医学图像分割 [J ] . 自动化学报 , 2023 , 49 ( 4 ): 805 - 819 .

FANG C W , LI X , LI Z Y , et al . Interactive dual-model learning for semi-supervised medical image segmentation [J ] . Acta Automatica Sinica , 2023 , 49 ( 4 ): 805 - 819 . (in Chinese)

YUN S , HAN D , CHUN S , et al . CutMix: Regularization strategy to train strong classifiers with localizable features [C ] // 2019 IEEE/CVF International Conference on Computer Vision (ICCV) . Piscataway : IEEE , 2019 : 6022 - 6031 .

OLSSON V , TRANHEDEN W , PINTO J , et al . ClassMix: Segmentation-based data augmentation for semi-supervised learning [C ] // 2021 IEEE Winter Conference on Applications of Computer Vision (WACV) . Piscataway : IEEE , 2021 : 1368 - 1377 .

BAI Y H , CHEN D W , LI Q L , et al . Bidirectional copy-paste for semi-supervised medical image segmentation [C ] // 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2023 : 11514 - 11524 .

OUALI Y , HUDELOT C , TAMI M . Semi-supervised semantic segmentation with cross-consistency training [C ] // 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2020 : 12674 - 12684 .

CHEN X K , YUAN Y H , ZENG G , et al . Semi-supervised semantic segmentation with cross pseudo supervision [C ] // 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2021 : 2613 - 2622 .

HUANG H J , CHEN Z G , CHEN C Y , et al . Complementary consistency semi-supervised learning for 3D left atrial image segmentation [J ] . Computers in Biology and Medicine , 2023 , 165 : 107368 .

CHI H Y , PANG J , ZHANG B F , et al . Adaptive bidirectional displacement for semi-supervised medical image segmentation [C ] // 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2024 : 4070 - 4080 .

FENG Z Y , ZHOU Q Y , GU Q Q , et al . DMT: Dynamic mutual training for semi-supervised learning [J ] . Pattern Recognition , 2022 , 130 : 108777 .

ZHAO Z , WANG Z C , WANG L Y , et al . Alternate Diverse Teaching for Semi-Supervised Medical Image Segmentation [M ] // Computer Vision-ECCV 2024 . Cham : Springer Nature Switzerland , 2024 : 227 - 243 .

HANG W L , BAI K , LIANG S , et al . Reliable semi-supervised mutual learning framework for medical image segmentation [J ] . Biomedical Signal Processing and Control , 2025 , 99 : 106798 .

ZHAO X Y , QI Z X , WANG S , et al . RCPS: Rectified contrastive pseudo supervision for semi-supervised medical image segmentation [J ] . IEEE Journal of Biomedical and Health Informatics , 2024 , 28 ( 1 ): 251 - 261 .

KENDALL A , GAL Y , SHRIDHAR K , et al . What uncertainties do we need in Bayesian deep learning for computer vision? [EB/OL ] . ( 2017-10-05 )[ 2024-07-02 ] . https://arxiv.org/abs/1703.04977v2 https://arxiv.org/abs/1703.04977v2 .

YU L Q , WANG S J , LI X M , et al . Uncertainty-aware self-ensembling model for semi-supervised 3D left atrium segmentation [M ] // Medical Image Computing and Computer Assisted Intervention-MICCAI 2019 . Cham : Springer International Publishing , 2019 : 605 - 613 .

LUO X D , WANG G T , LIAO W J , et al . Semi-supervised medical image segmentation via uncertainty rectified pyramid consistency [J ] . Medical Image Analysis , 2022 , 80 : 102517 .

WU Y C , GE Z Y , ZHANG D H , et al . Mutual consistency learning for semi-supervised medical image segmentation [J ] . Medical Image Analysis , 2022 , 81 : 102530 .

XIA Y D , LIU F Z , YANG D , et al . 3D semi-supervised learning with uncertainty-aware multi-view co-training [C ] // 2020 IEEE Winter Conference on Applications of Computer Vision (WACV) . Piscataway : IEEE , 2020 : 3646 - 3655 .

XIE Q Z , DAI Z H , HOVY E , et al . Unsupervised data augmentation for consistency training [EB/OL ] . ( 2020-11-05 )[ 2024-07-02 ] . https://arxiv.org/abs/1904.12848v6 https://arxiv.org/abs/1904.12848v6 .

LAINE S , AILA T . Temporal ensembling for semi-supervised learning [EB/OL ] . ( 2017-03-15 ) [ 2024-07-02 ] . https://doi.org/10.48550/arXiv.1610.02242 https://doi.org/10.48550/arXiv.1610.02242 .

WANG Y C , XIAO B , BI X L , et al . MCF: Mutual correction framework for semi-supervised medical image segmentation [C ] // 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE , 2023 : 15651 - 15660 .

XIONG Z H , XIA Q , HU Z Q , et al . A global benchmark of algorithms for segmenting the left atrium from late gadolinium-enhanced cardiac magnetic resonance imaging [J ] . Medical Image Analysis , 2021 , 67 : 101832 .

BAKAS S . Brats MICCAI Brain tumor dataset [DB/OL ] . IEEE Dataport , ( 2020-02-28 ) [ 2024-07-02 ] . https://ieee-dataport.org/competitions/brats-miccai-brain-tumor-dataset https://ieee-dataport.org/competitions/brats-miccai-brain-tumor-dataset .

XU Z , WANG Y X , LU D H , et al . All-around real label supervision: Cyclic prototype consistency learning for semi-supervised medical image segmentation [J ] . IEEE Journal of Biomedical and Health Informatics , 2022 , 26 ( 7 ): 3174 - 3184 .

BERNARD O , LALANDE A , ZOTTI C , et al . Deep learning techniques for automatic MRI cardiac multi-structures segmentation and diagnosis: Is the problem solved? [J ] . IEEE Transactions on Medical Imaging , 2018 , 37 ( 11 ): 2514 - 2525 .

LUO X D , HU M H , SONG T , et al . Semi-supervised medical image segmentation via cross teaching between CNN and transformer [C ] // International Conference on Medical Imaging with Deep Learning . New York : PMLR , 2022 : 820 - 833 .

MILLETARI F , NAVAB N , AHMADI S A . V-net: Fully convolutional neural networks for volumetric medical image segmentation [C ] // 2016 Fourth International Conference on 3D Vision (3DV) . Piscataway : IEEE , 2016 : 565 - 571 .

LUO X D , CHEN J N , SONG T , et al . Semi-supervised medical image segmentation through dual-task consistency [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2021 , 35 ( 10 ): 8801 - 8809 .

WU Y C , WU Z H , WU Q Y , et al . Exploring smoothness and class-separation for semi-supervised medical image segmentation [M ] // Medical Image Computing and Computer Assisted Intervention - MICCAI 2022 . Cham : Springer Nature Switzerland , 2022 : 34 - 43 .

XU Z , WANG Y X , LU D H , et al . Ambiguity-selective consistency regularization for mean-teacher semi-supervised medical image segmentation [J ] . Medical Image Analysis , 2023 , 88 : 102880 .

ZHANG Y C , JIAO R S , LIAO Q C , et al . Uncertainty-guided mutual consistency learning for semi-supervised medical image segmentation [J ] . Artificial Intelligence in Medicine , 2023 , 138 : 102476 .

ADIGA V S , DOLZ J , LOMBAERT H . Anatomically-aware uncertainty for semi-supervised image segmentation [J ] . Medical Image Analysis , 2024 , 91 : 103011 .

CHEN Y Y , YANG Z Y , SHEN C Y , et al . Evidence-based uncertainty-aware semi-supervised medical image segmentation [J ] . Computers in Biology and Medicine , 2024 , 170 : 108004 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于交叉视觉状态空间与多分支交互注意力的医学图像分割

基于混洗特征编码与门控解码的医学图像分割网络

基于联邦学习的主动半监督短文本分类方法