Research map / Point clouds / Point cloud classification
Perturbation Robust Representations of Topological Persistence Diagrams
Anirudh Som, Kowshik Thopalli, Karthikeyan Natesan Ramamurthy, Vinay Venkataraman, Ankita Shukla, Pavan Turaga
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 |
|---|---|---|---|
| fuse PTS-D3M features with SSM-HOG descriptor | SSM-HOG · IXMAS same camera | accuracy (%): 67.30 -> 76.18 | Improved |
| fuse PTS-D3M features with SSM-OF descriptor | SSM-OF · IXMAS same camera | accuracy (%): 66.60 -> 74.41 | Improved |
| fuse PTS-D3M features with SSM-HOG-OF descriptor | SSM-HOG-OF · IXMAS same camera | accuracy (%): 76.28 -> 81.12 | Improved |
| fuse PTS-D3M features with SSM-HOG-OF descriptor | SSM-HOG-OF · IXMAS any-to-any | accuracy (%): 61.25 -> 66.16 | Improved |
| use PTS computed over HOG descriptor fused with SSM-HOG instead of multi-scale shape distributions (ablation) | SSM-HOG · IXMAS same camera | accuracy (%): 67.30 -> 69.01 | Improved |
| use D3M shape distribution instead of A3M for PTS (ablation) | SSM-HOG-OF + PTS · IXMAS same camera | accuracy (%): 79.30 -> 81.12 | Improved |
| use PTS with d_Delta instead of 2-Wasserstein distance on PDs | PD representation · SHREC 2010 noisy shapes | average accuracy (%): 95.00 -> 99.03 | Improved |
| use PTS (d_Delta) instead of persistence images (L1) | PI · SHREC 2010 noisy shapes | average accuracy (%): 84.39 -> 99.03 | Improved |
| use PTS (d_Delta) instead of 1-Wasserstein distance on PDs | PD · SHREC 2010 noisy shapes | average accuracy (%): 98.93 -> 99.03 | No clear effect |
| use PTS (d_Delta) instead of 1-Wasserstein distance on PDs | PD · SHREC 2010 noisy shapes | average time per comparison (1e-4 s): 256.00 -> 1.60 | Improved |
| use d_Delta instead of d_G metric for PTS (ablation) | PTS · SHREC 2010 noisy shapes | average accuracy (%): 98.89 -> 99.03 | No clear effect |
| use PTS (d_Delta) NN instead of 1-Wasserstein PD NN | PD · Motion capture | accuracy (%): 93.68 -> 85.96 | Hurt |
Closest papers by meaning
- Learning Hyperbolic Representations of Topological Features (ICLR 2021 · similarity 0.90)
- Deep Learning with Topological Signatures (NeurIPS 2017 · similarity 0.89)
- A Framework for Fast and Stable Representations of Multiparameter Persistent Homology Decompositions (NeurIPS 2023 · similarity 0.88)
- Towards Scalable Topological Regularizers (ICLR 2025 · similarity 0.88)
- Point-Level Topological Representation Learning on Point Clouds (ICML 2025 · similarity 0.87)
- Diffeomorphic interpolation for efficient persistence-based topological optimization (NeurIPS 2024 · similarity 0.87)
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.
