EUVIMEDEUVIMED entdecken
Crossref

Text-Image Co-Alignment for Weakly Supervised Polyp Segmentation

Wenhui Huang, Zhen Pan, Xiaoyan Wang, Yedi Zhang, Jingzhen He, James C Gee, Yuanjie Zheng

IEEE Transactions on Medical Imaging · 2026-06 · 45 (6) · 3394-3408

DOI: 10.1109/tmi.2026.3674592 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Medical Imaging, 2026-06. Verfasst von Wenhui Huang und 6 weiteren Autor:innen. Hauptthemen laut MEDLINE: Polyps, Image Interpretation, Computer-Assisted, Supervised Machine Learning, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 15-mal zitiert (Crossref) und nennt 57 Literaturangaben. Gefördert von: Funds from the National Natural Science Foundation of China, Shandong Provincial Natural Science Foundation under, Taishan Scholars Youth Expert Program of Shandong Province.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Fully supervised polyp segmentation relies on costly pixel-level annotations. Although semi- and weakly supervised methods reduce annotation requirements, they still depend on partial mask supervision. Text-supervised segmentation is a promising alternative; however, for polyps, the key challenge is to ground instance-specific phrases to the correct lesion region under cluttered backgrounds and large appearance variations. Existing approaches often rely on coarse text-image alignment, limiting precise region-level semantic correspondence. In this paper, we propose Text-Image Co-Alignment (TICoA), a text-supervised framework for polyp segmentation. TICoA leverages large language models (LLMs)-generated structured clinical descriptions as weak supervision and formulates segmentation as a fine-grained phrase-region co-alignment problem. Through contrastive learning, TICoA explicitly associates query phrases with corresponding image regions to achieve robust semantic grounding under weak supervision. Architecturally, we adopt a State-Space Model (Mamba) to efficiently model long-range dependencies with linear computational complexity. To support effective cross-modal interaction, we further design a dedicated Mamba Fusion module with a Bi-Dimension Fusion (BiDF) strategy, which progressively propagates information along spatial and channel dimensions. Experiments on polyp datasets, with additional validation on skin lesion segmentation, demonstrate that TICoA is competitive with state-of-the-art weakly supervised methods. Our code and data are available at https://github.com/silentyuchen/TICoA.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

PolypsImage Interpretation, Computer-AssistedSupervised Machine LearningImage Processing, Computer-AssistedHumansAlgorithmsLarge Language Models

Förderung

  • Funds from the National Natural Science Foundation of China
  • Shandong Provincial Natural Science Foundation under
  • Taishan Scholars Youth Expert Program of Shandong Province
  • Young Talent of Lifting Engineering for Science and Technology in Shandong
  • Youth Innovation Technology Project of Higher Education in Shandong Province
  • Youth Innovation Technology Project of Higher Education in Shandong Province under

Zitieren

APA

Huang, W., Pan, Z., Wang, X., Zhang, Y., He, J., Gee, J. C., & Zheng, Y. (2026). Text-Image Co-Alignment for Weakly Supervised Polyp Segmentation. IEEE Transactions on Medical Imaging, 45(6), 3394-3408. https://doi.org/10.1109/tmi.2026.3674592

Vancouver

Huang W, Pan Z, Wang X, Zhang Y, He J, Gee JC, et al. Text-Image Co-Alignment for Weakly Supervised Polyp Segmentation. IEEE Transactions on Medical Imaging. 2026;45(6):3394-3408. doi:10.1109/tmi.2026.3674592

BibTeX
@article{Huang2026,
  author = {Huang, Wenhui and Pan, Zhen and Wang, Xiaoyan and Zhang, Yedi and He, Jingzhen and Gee, James C and Zheng, Yuanjie},
  title = {Text-Image Co-Alignment for Weakly Supervised Polyp Segmentation},
  journal = {IEEE Transactions on Medical Imaging},
  year = {2026},
  volume = {45},
  number = {6},
  pages = {3394--3408},
  doi = {10.1109/tmi.2026.3674592}
}
Kostenlos starten →