Research map / Probabilistic and scientific modeling / Molecules and proteins
Learning Probabilistic Topological Representations Using Discrete Morse Theory
Xiaoling Hu, Dimitris Samaras, Chao Chen
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 |
|---|---|---|---|
| Predict structures using a learned probabilistic representation built from discrete Morse theory and persistent homology | Biomedical image segmentation · ISBI13 | ARI: DMT: 0.8289 $\pm$ 0.0189; Ours: 0.8417 $\pm$ 0.0114 | Improved |
| Predict structures using a learned probabilistic representation built from discrete Morse theory and persistent homology | Biomedical image segmentation · ISBI13 | Dice: DMT: 0.9712 $\pm$ 0.0047; Ours: 0.9637 $\pm$ 0.0032 | Hurt |
| Predict structures using a learned probabilistic representation built from discrete Morse theory and persistent homology | Biomedical image segmentation · CREMI | Dice: DMT: 0.9653 $\pm$ 0.0019; Ours: 0.9681 $\pm$ 0.0016 | Improved |
| Predict structures using a learned probabilistic representation built from discrete Morse theory and persistent homology | Biomedical image segmentation · CREMI | ARI: DMT: 0.8203 $\pm$ 0.0147; Ours: 0.8475 $\pm$ 0.0043 | Improved |
| Predict structures using a learned probabilistic representation built from discrete Morse theory and persistent homology | Biomedical image segmentation · DRIVE | Dice: DMT: 0.7733 $\pm$ 0.0039; Ours: 0.7814 $\pm$ 0.0026 | Improved |
| Predict structures using a learned probabilistic representation built from discrete Morse theory and persistent homology | Biomedical image segmentation · DRIVE | VOI: DMT: 0.876 $\pm$ 0.038; Ours: 0.804 $\pm$ 0.047 | Improved |
| Jointly train the structure-aware model instead of applying Morse postprocessing to a standard network (ablation) | Proposed model · DRIVE | VOI: 1.165 $\pm$ 0.086 -> 0.804 $\pm$ 0.047 | Improved |
| Jointly train the structure-aware model instead of applying Morse postprocessing to a standard network (ablation) | Proposed model · DRIVE | Dice: 0.7653 $\pm$ 0.0052 -> 0.7814 $\pm$ 0.0026 | Improved |
| Set the KL divergence loss weight to $\beta=10$ (ablation) | Proposed model · DRIVE | VOI: 0.804 $\pm$ 0.047 | Improved |
Closest papers by meaning
- Topology-Aware Segmentation Using Discrete Morse Theory (ICLR 2021 · similarity 0.91)
- Topology-Aware Uncertainty for Image Segmentation (NeurIPS 2023 · similarity 0.90)
- Scale-Free Image Keypoints Using Differentiable Persistent Homology (ICML 2024 · similarity 0.87)
- Point-Level Topological Representation Learning on Point Clouds (ICML 2025 · similarity 0.86)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.86)
- Graphcode: Learning from multiparameter persistent homology using graph neural networks (NeurIPS 2024 · similarity 0.85)
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.
