Research map / Point clouds / Point cloud classification
On the Effectiveness of Persistent Homology
Read the paperCheck 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 persistent-homology features instead of distance-matrix SVMs, fully connected networks, or PointNet | Shape classification by number of holes · Synthetic 2D and 3D point clouds | test accuracy | Helped |
| Use persistent-homology features instead of the evaluated baselines | Curvature prediction · Synthetic point clouds sampled from constant-curvature surfaces | mean squared error | Helped |
| Use persistent-homology features instead of the evaluated baselines | Convexity classification · Synthetic regular and random shapes, including out-of-distribution tests | classification accuracy | Helped |
| Use persistent-homology features instead of PointNet | Convexity classification · Regular shapes | classification accuracy | No clear effect |
| Use all persistence-interval lifespans instead of only the 10 longest (ablation) | PH curvature predictor · Curvature prediction | mean squared error | Helped |
| Use only the 10 longest persistence-interval lifespans instead of all intervals (ablation) | PH curvature predictor · Curvature prediction | mean squared error | Hurt |
| Use the simple PH signature of all 0-dimensional lifespans without grid-search tuning (ablation) | PH curvature predictor · Curvature prediction | mean squared error | No clear effect |
| Use the simple PH signature of the 10 longest lifespans without PH-parameter tuning (ablation) | PH hole-count classifier · Number of holes | test accuracy | No clear effect |
| Use PH features with limited training data instead of the evaluated baseline pipelines | Shape classification by number of holes · Limited-training-data tests | test accuracy | Helped |
| Use PH features on transformed or noisy test point clouds instead of the evaluated baseline pipelines | Shape classification by number of holes · Affine transformations and noise | test accuracy | Helped |
| Use PH features rather than PointNet | Shape classification by number of holes · Translated or rotated test data and tests with outliers | test accuracy | Helped |
Closest papers by meaning
- Point-Level Topological Representation Learning on Point Clouds (ICML 2025 · similarity 0.90)
- The Flood Complex: Large-Scale Persistent Homology on Millions of Points (NeurIPS 2025 · similarity 0.90)
- Scale-Free Image Keypoints Using Differentiable Persistent Homology (ICML 2024 · similarity 0.87)
- Persistent Homology for High-dimensional Data Based on Spectral Methods (NeurIPS 2024 · similarity 0.87)
- Towards Scalable Topological Regularizers (ICLR 2025 · similarity 0.86)
- Wasserstein convergence of Cech persistence diagrams for samplings of submanifolds (NeurIPS 2024 · similarity 0.86)
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.