DTU-Net: Learning Topological Similarity for Curvilinear Structure Segmentation
This addresses segmentation challenges in medical imaging for structures like vessels and neurons, offering a more efficient alternative to prior topological methods, though it is incremental as it builds on U-Net architectures.
The paper tackles the problem of segmenting small, low-contrast curvilinear structures in medical imaging by introducing DTU-Net, a data-driven method that uses two sequential U-Nets to preserve topology without prior knowledge, achieving improved pixel-wise accuracy and topological continuity on ultrasound and retinal datasets.
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.