Research map / Detection and segmentation / CNN features
Progressive Minimal Path Method With Embedded CNN
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use Path-CNN instead of the strongest baseline, PCNN | Centerline segmentation · EPFL | mean distance (pixels): 2.74 -> 1.68 | Helped |
| Use Path-CNN instead of the strongest baseline, PCNN | Centerline segmentation · Road | mean distance (pixels): 3.68 -> 2.97 | Helped |
| Use Path-CNN instead of the strongest baseline, PCNN | Centerline segmentation · River | mean distance (pixels): 6.31 -> 3.78 | Helped |
| Use Path-CNN instead of the strongest baseline, PCNN | Centerline segmentation · Toronto | mean distance (pixels): 5.25 -> 3.59 | Helped |
| Train U-Net with centerline annotations instead of Path-CNN | EPFL road mask segmentation · EPFL | Dice coefficient: 0.6108 -> 0.1163 | Hurt |
| Train U-Net with mask annotations instead of Path-CNN | EPFL road mask segmentation · EPFL | Dice coefficient: 0.6108 -> 0.2953 | Hurt |
| Embed an axis-aligned-patch classifier in the minimal-path method instead of classifying pixels independently (ablation) | Pixel classification and minimal-path framework · EPFL road F-map | classification quality | Helped |
| Train the path classifier with rectified patches instead of axis-aligned patches (ablation) | Minimal-path framework with CNN classification · EPFL road F-map | classification quality | Helped |
| Increase image patch size (ablation) | CNN patch classifier · EPFL classification | precision | Mixed |
| Change CNN architecture among the tested architectures (ablation) | Rectified-patch CNN classifier · EPFL classification | classification performance | No clear effect |
Closest papers by meaning
- clDice - A Novel Topology-Preserving Loss Function for Tubular Structure Segmentation (CVPR 2021 · similarity 0.87)
- GraphMorph: Tubular Structure Extraction by Morphing Predicted Graphs (NeurIPS 2024 · similarity 0.85)
- SplineCNN: Fast Geometric Deep Learning With Continuous B-Spline Kernels (CVPR 2018 · similarity 0.85)
- One-dimensional Path Convolution (ICML 2025 · similarity 0.84)
- SegGraph: Leveraging Graphs of SAM Segments for Few-Shot 3D Part Segmentation (NeurIPS 2025 · similarity 0.84)
- Upping the Game: How 2D U-Net Skip Connections Flip 3D Segmentation (NeurIPS 2024 · similarity 0.83)
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.