Research map / Detection and segmentation / CNN features
NETracer: A Topology-Aware Iterative Tracing Approach for Tubular Structure Extraction
Chao Liu, Yangbo Jiang, Nenggan Zheng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add edge prediction loss Le (ablation) | M1 node-only tracer · fMOST-VTA (3D) | Jump Error: 36 -> 32 | Improved |
| add edge prediction loss Le (ablation) | M1 node-only tracer · fMOST-VTA (3D) | SSD-F1: 0.892 -> 0.897 | Improved |
| add local connectivity loss Lconn (ablation) | M2 with edge loss · fMOST-VTA (3D) | Jump Error: 32 -> 26 | Improved |
| add local connectivity loss Lconn (ablation) | M2 with edge loss · fMOST-VTA (3D) | SSD-F1: 0.897 -> 0.903 | Improved |
| add geodesic distance-based search (GDS) with edge cues (ablation) | M4 · fMOST-VTA (3D) | Jump Error: 26 -> 18 | Improved |
| add geodesic distance-based search (GDS) with edge cues (ablation) | M4 · fMOST-VTA (3D) | SSD-F1: 0.903 -> 0.919 | Improved |
| apply GDS directly on predicted centerline without edge cues (ablation) | M1 · fMOST-VTA (3D) | Jump Error: 36 -> 51 | Hurt |
| apply GDS directly on predicted centerline without edge cues (ablation) | M1 · fMOST-VTA (3D) | SSD-F1: 0.892 -> 0.893 | Hurt |
| use NETracer with GDS instead of DBT (ablation) | DBT · fMOST-VTA (3D) | Total runtime: 278.4s -> 97.8s | Improved |
| use NETracer instead of Vec-Road (ablation) | Vec-Road · fMOST-VTA (3D) | Total runtime: 84.5s -> 97.8s | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- RoadTracer — RoadTracer: Automatic Extraction of Road Networks From Aerial Images (CVPR 2018) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PointScatter: Point Set Representation for Tubular Structure Extraction (ECCV 2022 · similarity 0.88)
- Skeleton Recall Loss for Connectivity Conserving and Resource Efficient Segmentation of Thin Tubular Structures (ECCV 2024 · similarity 0.88)
- GraphMorph: Tubular Structure Extraction by Morphing Predicted Graphs (NeurIPS 2024 · similarity 0.88)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.84)
- TopoDiffusionNet: A Topology-aware Diffusion Model (ICLR 2025 · similarity 0.84)
- CurveBench: A Benchmark for Exact Topological Reasoning over Nested Jordan Curves (NeurIPS 2026 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
