Skip to content
2000
Volume 21, Issue 1
  • ISSN: 1573-4056
  • E-ISSN: 1875-6603
file format pdf download PDF
file format text download EPUB
side by side viewer icon HTML

Abstract

Introduction

This study aims to propose and evaluate a two-stage semi-supervised segmentation framework with dual multiscale uncertainty estimation and graph reasoning, addressing the challenges of obtaining high-precision pixel-level labels and effectively utilizing unlabeled data for accurate pneumonia lesion segmentation.

Methods

First, we design a guided supervised training strategy for modeling aleatoric uncertainty (AU) at dual scales, reducing the impact on segmentation performance caused by aleatoric uncertainties introduced by blurred lesions and their boundaries in the image. Second, we design a training strategy for multi-scale noisy pseudo-label correction to reduce the cognitive bias problem caused by unreliable predictions in the model. Finally, we design a new combination of fused feature interaction graph reasoning (FIGR) and attention modules, which enables the network model to better capture image features in small infected regions.

Results

Our study was validated using the MosMedData public dataset. The proposed algorithm improves the performance by 1.25%, 1.03%, 2.98%, and 0.59% on Dice, Jaccard, normalized surface dice (NSD), and average distance of boundaries (ADB), respectively, compared to the baseline model.

Discussion

Our semi-supervised pneumonia segmentation framework, through two-stage multi-scale uncertainty estimation and modeling, significantly improves segmentation performance by leveraging unlabeled data and addressing uncertainties, offering clinical benefits in pneumonia diagnosis while facing challenges in generalization and computational efficiency that future work will target with GAN-based data synthesis and architecture optimization.

Conclusion

It can be convincingly concluded that the proposed algorithm is of profound importance and value in the domain of clinical practice.

© 2025 The Author(s). Published by Bentham Science Publishers.
This is an open access article published under CC BY 4.0 https://creativecommons.org/licenses/by/4.0/legalcode
Loading

Article metrics loading...

/content/journals/cmir/10.2174/0115734056363870250804215311
2025-08-08
2025-10-29

  • From This Site

    /content/journals/cmir/10.2174/0115734056363870250804215311
    dcterms_title,dcterms_subject,pub_keyword
    -contentType:Contributor -contentType:Concept -contentType:Institution
    10
    5
Loading full text...

Full text loading...

/deliver/fulltext/cmir/21/1/CMIR-21-E15734056363870.html?itemId=/content/journals/cmir/10.2174/0115734056363870250804215311&mimeType=html&fmt=ahah

References

  1. HaiderN.S. BeheraA.K. Computerized respiratory sound based diagnosis of pneumonia.Med. Biol. Eng. Comput.20246219510610.1007/s11517‑023‑02935‑737723381
     [Google Scholar]
  2. TorresA. CillonizC. NiedermanM.S. MenéndezR. ChalmersJ.D. WunderinkR.G. van der PollT. Pneumonia.Nat. Rev. Dis. Primers2021712510.1038/s41572‑021‑00259‑033833230
     [Google Scholar]
  3. RajaramanS. CandemirS. KimI. ThomaG. AntaniS. Visualization and interpretation of convolutional neural network predictions in detecting pneumonia in pediatric chest radiographs.Appl. Sci.2018810171510.3390/app810171532457819
     [Google Scholar]
  4. AydoğduM. OzyilmazE. AksoyH. GürselG. EkimN. Mortality prediction in community-acquired pneumonia requiring mechanical ventilation; values of pneumonia and intensive care unit severity scores.Tuberk. Toraks2010581253420517726
     [Google Scholar]
  5. DongD. FangM.J. TangL. ShanX.H. GaoJ.B. GigantiF. WangR.P. ChenX. WangX.X. PalumboD. FuJ. LiW.C. LiJ. ZhongL.Z. De CobelliF. JiJ.F. LiuZ.Y. TianJ. Deep learning radiomic nomogram can predict the number of lymph node metastasis in locally advanced gastric cancer: An international multicenter study.Ann. Oncol.202031791292010.1016/j.annonc.2020.04.00332304748
     [Google Scholar]
  6. DongD. ZhangF. ZhongL.Z. FangM.J. HuangC.L. YaoJ.J. SunY. TianJ. MaJ. TangL.L. Development and validation of a novel MR imaging predictor of response to induction chemotherapy in locoregionally advanced nasopharyngeal cancer: A randomized controlled trial substudy (NCT01245959).BMC Med.201917119010.1186/s12916‑019‑1422‑631640711
     [Google Scholar]
  7. GuY. LuX. ZhangB. ZhaoY. YuD. GaoL. CuiG. WuL. ZhouT. Automatic lung nodule detection using multi-scale dot nodule-enhancement filter and weighted support vector machines in chest computed tomography.PLoS One2019141021055110.1371/journal.pone.021055130629724
     [Google Scholar]
  8. LakhaniP. SundaramB. Deep learning at chest radiography: Automated classification of pulmonary tuberculosis by using convolutional neural networks.Radiology2017284257458210.1148/radiol.201716232628436741
     [Google Scholar]
  9. LuX. GuY. YangL. ZhangB. ZhaoY. YuD. ZhaoJ. GaoL. ZhouT. LiuY. ZhangW. Multi-level 3D densenets for false-positive reduction in lung nodule detection based on chest computed tomography.Curr. Med. Imaging20201681004102110.2174/157340561566619111312284033081662
     [Google Scholar]
  10. WangB. LiM. MaH. HanF. WangY. ZhaoS. LiuZ. YuT. TianJ. DongD. PengY. Computed tomography-based predictive nomogram for differentiating primary progressive pulmonary tuberculosis from community-acquired pneumonia in children.BMC Med. Imaging20191916310.1186/s12880‑019‑0355‑z31395012
     [Google Scholar]
  11. GuY. ChiJ. LiuJ. YangL. ZhangB. YuD. ZhaoY. LuX. A survey of computer-aided diagnosis of lung nodules from CT scans using deep learning.Comput. Biol. Med.202113710480610.1016/j.compbiomed.2021.10480634461501
     [Google Scholar]
  12. MorozovS.P. AndreychenkoA.E. PavlovN.A. Mosmeddata: Chest ct scans with covid-19 related findings dataset.medRXiv20201710.1101/2020.05.20.20100362
     [Google Scholar]
  13. FengX. LinJ. FengC.M. LuG. GAN inversion-based semi-supervised learning for medical image segmentation.Biomed. Signal Process. Control202488105536[https://doi.org/10.1016/j.bspc.2023.105536].10.1016/j.bspc.2023.105536
     [Google Scholar]
  14. FanD.P. ZhouT. JiG.P. ZhouY. ChenG. FuH. ShenJ. ShaoL. Inf-net: Automatic covid-19 lung infection segmentation from ct images.IEEE Trans. Med. Imaging20203982626263710.1109/TMI.2020.299664532730213
     [Google Scholar]
  15. JiaoR. ZhangY. DingL. XueB. ZhangJ. CaiR. JinC. Learning with limited annotations: A survey on deep semi-supervised learning for medical image segmentation.Comput. Biol. Med.202416910784010.1016/j.compbiomed.2023.10784038157773
     [Google Scholar]
  16. TarvainenA. ValpolaH. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results.Adv. Neural Inf. Process. Syst.20173016
     [Google Scholar]
  17. YuL WangS LiX FuCW HengPA Uncertainty-aware self-ensembling model for semi-supervised 3D left atrium segmentation.Medical Image Computing and Computer Assisted Intervention – MICCAI 2019ChamSpringer11765201910.1007/978‑3‑030‑32245‑8_67
     [Google Scholar]
  18. LuL. YinM. FuL. YangF. Uncertainty-aware pseudo-label and consistency for semi-supervised medical image segmentation.Biomed. Signal Process. Control202379104203[https://doi.org/10.1016/j.bspc.2022.104203].10.1016/j.bspc.2022.104203
     [Google Scholar]
  19. XiangJ. QiuP. YangY. FUSSNet: Fusing two sources of uncertainty for semi-supervised medical image segmentation.Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 WangL DouQ FletcherPT SpeidelS LiS ChamSpringer13438202210.1007/978‑3‑031‑16452‑1_46
     [Google Scholar]
  20. LeiT. ZhangD. DuX. WangX. WanY. NandiA.K. Semi-supervised medical image segmentation using adversarial consistency learning and dynamic convolution network.IEEE Trans. Med. Imaging20234251265127710.1109/TMI.2022.322568736449588
     [Google Scholar]
  21. WuH. WangZ. SongY. YangL. QinJ. Cross-patch dense contrastive learning for semi-supervised segmentation of cellular nuclei in histopathologic images.Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition2022, pp. 11666-11675.10.1109/CVPR52688.2022.01137
     [Google Scholar]
  22. WangK. ZhanB. ZuC. WuX. ZhouJ. ZhouL. WangY. Semi-supervised medical image segmentation via a tripled-uncertainty guided mean teacher model with contrastive learning.Med. Image Anal.20227910244710.1016/j.media.2022.10244735509136
     [Google Scholar]
  23. WangL. WangJ. ZhuL. FuH. LiP. ChengG. FengZ. LiS. HengP.A. Dual multiscale mean teacher network for semi-supervised infection segmentation in chest CT volume for COVID-19.IEEE Trans. Cybern.202353106363637510.1109/TCYB.2022.322352837015538
     [Google Scholar]
  24. GuY. LuX. YangL. ZhangB. YuD. ZhaoY. GaoL. WuL. ZhouT. Automatic lung nodule detection using a 3D deep convolutional neural network combined with a multi-scale prediction strategy in chest CTs.Comput. Biol. Med.201810322023110.1016/j.compbiomed.2018.10.01130390571
     [Google Scholar]
  25. MonteiroM. Le FolgocL. Coelho de CastroD. PawlowskiN. MarquesB. KamnitsasK. van der WilkM. GlockerB. Stochastic segmentation networks: Modelling spatially correlated aleatoric uncertainty.Adv. Neural Inf. Process. Syst.2020331275612767
     [Google Scholar]
  26. WangY. XiaoB. BiX. LiW. GaoX. MCF: Mutual correction framework for semi-supervised medical image segmentation.Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition2023, pp. 15651-15660.10.1109/CVPR52729.2023.01502
     [Google Scholar]
  27. SohnK. BerthelotD. CarliniN. ZhangZ. ZhangH. RaffelC.A. CubukE.D. KurakinA. LiC.L. Fixmatch: Simplifying semi-supervised learning with consistency and confidence.Adv. Neural Inf. Process. Syst.202033596608
     [Google Scholar]
  28. LuoX LiaoW ChenJ SongT ChenY ZhangS ChenN WangG ZhangS Efficient semi-supervised gross target volume of nasopharyngeal carcinoma segmentation via uncertainty rectified pyramid consistency.Medical Image Computing and Computer Assisted Intervention – MICCAI 2021ChamSpringer12902202110.1007/978‑3‑030‑87196‑3_30
     [Google Scholar]
  29. ZhengZ. YangY. Rectifying pseudo label learning via uncertainty estimation for domain adaptive semantic segmentation.Int. J. Comput. Vis.2021129411061120[https://doi.org/10.1007/s11263-020-01395-y].10.1007/s11263‑020‑01395‑y
     [Google Scholar]
  30. ZhuangY. LiuH. SongE. HungC.C. A 3D cross-modality feature interaction network with volumetric feature alignment for brain tumor and tissue segmentation.IEEE J. Biomed. Health Inform.2023271758610.1109/JBHI.2022.321499936251915
     [Google Scholar]
  31. ChenY. RohrbachM. YanZ. ShuichengY. FengJ. KalantidisY. Graph-based global reasoning networks.2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)2019.10.1109/CVPR.2019.00052
     [Google Scholar]
  32. LiuZ. TongL. ChenL. ZhouF. JiangZ. ZhangQ. WangY. ShanC. LiL. ZhouH. Canet: Context aware network for brain glioma segmentation.IEEE Trans. Med. Imaging20214071763177710.1109/TMI.2021.306591833720830
     [Google Scholar]
  33. DosovitskiyA. BeyerL. KolesnikovA. An image is worth 16x16 words: Transformers for image recognition at scale.arXiv:2010.1192920201710.48550/arXiv.2010.11929
     [Google Scholar]
  34. LiuZ. LinY. CaoY. HuH. WeiY. ZhangZ. LinS. GuoB. Swin transformer: Hierarchical vision transformer using shifted windows.arXiv:2103.1403020211410.48550/arXiv.2103.14030
     [Google Scholar]
  35. IsenseeF. PetersenJ. KohlS.A. JägerP.F. Maier-HeinK.H. nnu-net: Breaking the spell on successful medical image segmentation.2019Available from: https://www.researchgate.net/publication/332494163_nnU-Net_Breaking_the_Spell_on_Successful_Medical_Image_Segmentation
  36. ÇiçekÖ AbdulkadirA LienkampSS BroxT RonnebergerO 3D U-Net: Learning dense volumetric segmentation from sparse annotation.Medical Image Computing and Computer-Assisted Intervention – MICCAI 2016 OurselinS JoskowiczL SabuncuM UnalG ChamSpringer9901201610.1007/978‑3‑319‑46723‑8_49
     [Google Scholar]
  37. MahmudT. AlamM.J. ChowdhuryS. AliS.N. RahmanM.M. Anowarul FattahS. SaquibM. CovTANet: A hybrid tri-level attention-based network for lesion segmentation, diagnosis, and severity prediction of COVID-19 chest CT scans.IEEE Trans. Industr. Inform.20211796489649810.1109/TII.2020.304839137981913
     [Google Scholar]
  38. XuZ. CaoY. JinC. ShaoG. LiuX. ZhouJ. ShiH. FengJ.J. Gasnet: Weakly-supervised framework for covid-19 lesion segmentation.2020Available from: https://www.researchgate.net/publication/344757137_GASNet_Weakly-supervised_Framework_for_COVID-19_Lesion_Segmentation
  39. ShabaniS. HomayounfarM. VardhanabhutiV. Nikouei MahaniM.A. Koohi-MoghadamM. Self-supervised region-aware segmentation of COVID-19 CT images using 3D GAN and contrastive learning.Comput. Biol. Med.202214910603310.1016/j.compbiomed.2022.10603336041270
     [Google Scholar]
  40. ZhouZ. Rahman SiddiqueeM.M. TajbakhshN. LiangJ. UNet++: A nested U-Net architecture for medical image segmentation.Proceedings of the 4th International Workshop2018, pp. 3-11.10.1007/978‑3‑030‑00889‑5_1
     [Google Scholar]
  41. MaJ. WangY. AnX. GeC. YuZ. ChenJ. ZhuQ. DongG. HeJ. HeZ. CaoT. ZhuY. NieZ. YangX. Toward data‐efficient learning: A benchmark for COVID‐19 CT lung and infection segmentation.Med. Phys.20214831197121010.1002/mp.1467633354790
     [Google Scholar]
  42. AlhussenA. Anul HaqM. Ahmad KhanA. MahendranR.K. KadryS. XAI-RACapsNet: Relevance aware capsule network-based breast cancer detection using mammography images via explainability O-net ROI segmentation.Expert Syst. Appl.2025261125461[https://doi.org/10.1016/j.eswa.2024.125461].10.1016/j.eswa.2024.125461
     [Google Scholar]
  43. KhanA.A. MadendranR.K. ThirunavukkarasuU. FaheemM. D 2 PAM : Epileptic seizures prediction using adversarial deep dual patch attention mechanism.CAAI Trans. Intell. Technol.202383755769[https://doi.org/10.1049/cit2.12261].10.1049/cit2.12261
     [Google Scholar]
  44. KujurA RazaZ KhanAA Data complexity based evaluation of the model dependence of brain MRI images for classification of brain tumor and Alzheimer’s disease.IEEE Access2022101510.1109/ACCESS.2022.3216393
     [Google Scholar]
/content/journals/cmir/10.2174/0115734056363870250804215311
Loading
GRMA-Net: A novel two-stage 3D Semi-supervised Pneumonia Segmentation based on Dual Multiscale Uncertainty Estimation with Graph Reasoning in Chest CTs
Curr. Med. Imaging 21, E15734056363870 (2025); https://doi.org/10.2174/0115734056363870250804215311
/content/journals/cmir/10.2174/0115734056363870250804215311
/content/journals/cmir/10.2174/0115734056363870250804215311
Loading

Data & Media loading...


  • Article Type:     Research Article
Keyword(s): Graph reasoning; Mean teacher; Medical image segmentation; NSD; Semi-supervised Learning; Uncertainty estimate

Most Cited Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error
Please enter a valid_number test