EUVIMEDDescoperă EUVIMED
Open access · CC BY 4.0

TSSP‐UNet: A Two‐Stage Weakly Supervised Pathological Image Segmentation With Point Annotations

Shaoqiang Wang iD, Guiling Shi, Yuchen Wang, Qiang Li, Yawu Zhao, Xiaochun Cheng

IET Systems Biology · 2026-01 · 20 (1)

DOI: 10.1049/syb2.70055 ↗

At a glance

Journal article in IET Systems Biology, published online on 2026-03-03. Written by Shaoqiang Wang and 5 co-authors. Institution: Qingdao University of Technology, Qingdao, China. Major MEDLINE topics: Image Processing, Computer-Assisted, Image Interpretation, Computer-Assisted. 34 references. The full text is freely available (open access, CC BY 4.0). Funded by: UKRI, Doctoral Training Centre at Swansea University.

RezumatSubjectsDate bibliograficeCiteRelated studies

Abstract (original language)

Deep convolutional neural networks have demonstrated remarkable effectiveness in image segmentation. However, segmentation becomes challenging when training on images with complex instances. Moreover, obtaining annotations for high‐precision data is also difficult. Weakly supervised learning can address this issue by using nonspecialised annotations or supervised information from segmentation algorithms. In this study, we proposed TSSP‐UNet: a two‐stage weakly supervised segmentation approach. In the first stage, we trained a segmentation network augmented with constraint and attention mechanisms. These mechanisms are designed to operate on boundaries and superpixels generated from pseudo‐labels. For the attention network, two pseudo‐labels were used with a binary mask to add contour information to the segmentation process. Furthermore, a feature aggregation segmentation network was applied to the prominent foreground area in the image by incrementally adding elements. In the second stage, a refined confident learning algorithm improved the pseudo‐labels at the pixel level and then TSSP‐UNet was retrained using the modified superpixel labels. Testing on the MoNuSeg and TNBC datasets demonstrates that the approach performs well in the weakly supervised cell nucleus segmentation task compared with baseline methods.

Subjects

MeSH terms (MEDLINE)

Image Processing, Computer-AssistedImage Interpretation, Computer-AssistedHumansConvolutional Neural NetworksSupervised Machine LearningAlgorithms

Author keywords

image segmentationmachine learningneural network

Authors and institutions

  • Qingdao University of Technology, Qingdao, China.
  • Peking University People's Hospital, Qingdao, China.
  • Shandong University of Traditional Chinese Medicine, Jinan, China.
  • Computer Science Department, Bay Campus, Fabian Way, Swansea University, Swansea, UK.

ORCID profiles: Shaoqiang Wang

Funding

  • UKRI
  • Doctoral Training Centre at Swansea University

Cite

APA

Wang, S., Shi, G., Wang, Y., Li, Q., Zhao, Y., & Cheng, X. (2026). TSSP‐UNet: A Two‐Stage Weakly Supervised Pathological Image Segmentation With Point Annotations. IET Systems Biology, 20(1). https://doi.org/10.1049/syb2.70055

Vancouver

Wang S, Shi G, Wang Y, Li Q, Zhao Y, Cheng X. TSSP‐UNet: A Two‐Stage Weakly Supervised Pathological Image Segmentation With Point Annotations. IET Systems Biology. 2026;20(1). doi:10.1049/syb2.70055

BibTeX
@article{Wang2026,
  author = {Wang, Shaoqiang and Shi, Guiling and Wang, Yuchen and Li, Qiang and Zhao, Yawu and Cheng, Xiaochun},
  title = {TSSP‐UNet: A Two‐Stage Weakly Supervised Pathological Image Segmentation With Point Annotations},
  journal = {IET Systems Biology},
  year = {2026},
  volume = {20},
  number = {1},
  doi = {10.1049/syb2.70055}
}
Start for free →