DTU-Net: Learning Topological Similarity for Curvilinear Structure Segmentation

Manxi Lin, Kilian Zepf, Anders Nymark Christensen, Zahra Bashir, Morten Bo Søndergaard Svendsen, Martin Tolsgaard, Aasa Feragen*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review


Curvilinear structure segmentation is important in medical imaging, quantifying structures such as vessels, airways, neurons, or organ boundaries in 2D slices. Segmentation via pixel-wise classification often fails to capture the small and low-contrast curvilinear structures. Prior topological information is typically used to address this problem, often at an expensive computational cost, and sometimes requiring prior knowledge of the expected topology. 

We present DTU-Net, a data-driven approach to topology-preserving curvilinear structure segmentation. DTU-Net consists of two sequential, lightweight U-Nets, dedicated to texture and topology, respectively. While the texture net makes a coarse prediction using image texture information, the topology net learns topological information from the coarse prediction by employing a triplet loss trained to recognize false and missed splits in the structure. We conduct experiments on a challenging multi-class ultrasound scan segmentation dataset as well as a well-known retinal imaging dataset. Results show that our model outperforms existing approaches in both pixel-wise segmentation accuracy and topological continuity, with no need for prior topological knowledge.

Original languageEnglish
Title of host publicationProceedings of the 28th International Conference of Information Processing in Medical Imaging, IPMI 2023
Publication date2023
ISBN (Print)978-3-031-34047-5
ISBN (Electronic)978-3-031-34048-2
Publication statusPublished - 2023
Event28th International Conference on Information Processing in Medical Imaging, IPMI 2023 - San Carlos de Bariloche, Argentina
Duration: 18 Jun 202323 Jun 2023


Conference28th International Conference on Information Processing in Medical Imaging, IPMI 2023
CitySan Carlos de Bariloche


  • Curvilinear segmentation
  • Topology preservation
  • Triplet loss


Dive into the research topics of 'DTU-Net: Learning Topological Similarity for Curvilinear Structure Segmentation'. Together they form a unique fingerprint.

Cite this