EUVIMEDEUVIMED entdecken
Crossref

Adversarial consistency-based semi-supervised pneumonia segmentation using dual multiscale feature selection and fusion mean teacher model and triple-attention dynamic convolution in chest CTs

Yu Gu iD, Jianning Zang, Lidong Yang iD, Baohua Zhang, Jing Wang, Xiaoqi Lu, Jianjun Li, Xin Liu iD, Ying Zhao iD, Dahua Yu, Siyuan Tang, Qun He

Journal of X-Ray Science and Technology · 2025

DOI: 10.1177/08953996251367210 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of X-Ray Science and Technology, 2025. Verfasst von Yu Gu und 11 weiteren Autor:innen. Beteiligte Einrichtung: School of Digital and Intelligent Industry, Inner Mongolia University of Science and Technology, Baotou, China. Hauptthemen laut MEDLINE: Tomography, X-Ray Computed, Pneumonia, Image Processing, Computer-Assisted, Supervised Machine Learning. Die Arbeit nennt 53 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Recently, semi-supervised learning has demonstrated significant potential in the field of medical image segmentation. However, the majority of the methods fail to establish connections among diverse sample data. Moreover, segmentation networks that utilize fixed parameters can impede model training and even amplify the risk of overfitting. To address these challenges, this paper proposes an adversarial consistency-based semi-supervised segmentation method, leveraging a dual multiscale mean teacher model. First, by designing a discriminator network with adaptive feature selection and training it alternately with the segmentation network, the method enhances the segmentation network's ability to transfer knowledge from the limited labeled data to the unlabeled data. The discriminator evaluates the quality of the segmentation network's results for both labeled and unlabeled data, while simultaneously guiding the network to learn consistency in segmentation performance throughout the training process. Second, we design a Triple-attention dynamic convolutional (TADC) module, which allows the convolution kernel parameters to be adjusted flexibly according to different input data. This improves the feature representation capability of the network model and helps reduce the risk of overfitting. Finally, we propose a novel feature selection and fusion module (FSFM) within the segmentation network, which dynamically selects and integrates important features to enhance the saliency of key information, improving the overall performance of the model. The proposed adversarial consistency-based semi-supervised segmentation method is applied to the MosMedData dataset. The results demonstrate that the segmentation network outperforms the baseline model, achieving improvements of 3.83%, 3.97%, 3.14% in terms of Dice, Jaccard, and NSD scores, respectively, for the segmentation of pneumonia lesions. The proposed segmentation method outperforms state-of-the-art segmentation networks and demonstrates superior potential for segmenting pneumonia lesions, as evidenced by extensive experiments conducted on the MosMedData and COVID-19-P20 datasets.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Tomography, X-Ray ComputedPneumoniaImage Processing, Computer-AssistedSupervised Machine LearningHumansCOVID-19AlgorithmsLungNeural Networks, ComputerDeep Learning

Stichwörter der Autor:innen

adversarial consistency learningdynamic convolutionmean teachermedical image segmentationsemi-supervised learning

Autor:innen und Institutionen

  • School of Digital and Intelligent Industry, Inner Mongolia University of Science and Technology, Baotou, China.
  • School of Automation and Electrical Engineering, Inner Mongolia University of Science and Technology, Baotou, China.
  • School of Information and Electronics, Beijing Institute of Technology, Beijing, China.
  • College of Information Engineering, Inner Mongolia University of Technology, Hohhot, China.
  • School of Computer Science and Technology, Baotou Medical College, Inner Mongolia University of Science and Technology, Baotou, China.

ORCID-Profile: Yu Gu, Lidong Yang, Xin Liu, Ying Zhao

Zitieren

APA

Gu, Y., Zang, J., Yang, L., Zhang, B., Wang, J., Lu, X., Li, J., Liu, X., Zhao, Y., Yu, D., Tang, S., & He, Q. (2025). Adversarial consistency-based semi-supervised pneumonia segmentation using dual multiscale feature selection and fusion mean teacher model and triple-attention dynamic convolution in chest CTs. Journal of X-Ray Science and Technology. https://doi.org/10.1177/08953996251367210

Vancouver

Gu Y, Zang J, Yang L, Zhang B, Wang J, Lu X, et al. Adversarial consistency-based semi-supervised pneumonia segmentation using dual multiscale feature selection and fusion mean teacher model and triple-attention dynamic convolution in chest CTs. Journal of X-Ray Science and Technology. 2025. doi:10.1177/08953996251367210

BibTeX
@article{Gu2025,
  author = {Gu, Yu and Zang, Jianning and Yang, Lidong and Zhang, Baohua and Wang, Jing and Lu, Xiaoqi and Li, Jianjun and Liu, Xin and Zhao, Ying and Yu, Dahua and Tang, Siyuan and He, Qun},
  title = {Adversarial consistency-based semi-supervised pneumonia segmentation using dual multiscale feature selection and fusion mean teacher model and triple-attention dynamic convolution in chest CTs},
  journal = {Journal of X-Ray Science and Technology},
  year = {2025},
  doi = {10.1177/08953996251367210}
}
Kostenlos starten →