Research map / Detection and segmentation / CNN features
CuMPerLay: Learning Cubical Multiparameter Persistence Vectorizations
Caner Korkmaz, Brighton Nuwagira, Baris Coskunuzer, Tolga Birdal
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 learnable cubical multipersistence (TopoSwin-MP) to Swin | SwinV2-B · Glaucoma | AUC: 92.24 -> 92.87 | Improved |
| add learnable cubical multipersistence (TopoSwin-MP) to Swin | SwinV2-B · DDSM | AUC: 74.17 -> 76.32 | Improved |
| add learnable cubical multipersistence (TopoSwin-MP) to Swin | SwinV2-B · ISIC | AUC: 95.54 -> 96.50 | Improved |
| add learnable cubical multipersistence (TopoSwin-MP) to Swin | SwinV2-B · ISIC | Accuracy: 84.39 -> 83.47 | Hurt |
| use TopoSwin-MP instead of PHG-Net | SwinV2-B · DDSM | AUC: 61.07 -> 76.32 | Improved |
| use multiparameter instead of single persistence in standalone vectors (ablation) | MLP on topological vectors · Glaucoma | AUC: 81.48 -> 83.22 | Improved |
| add MP Betti vectors with traditional filtrations to Swin (ablation) | SwinV2-B · DDSM | AUC: 74.17 -> 72.58 | Hurt |
| add Topo-MP to U-Net skip connections | U-Net (from scratch) · VOC 2012 | mIoU: 23.96 -> 30.42 | Improved |
| add Topo-MP to FCN | FCN ResNet-50 · VOC 2012 | mIoU: 38.73 -> 40.36 | Improved |
| use Topo-MP instead of Topo-SP on U-Net (ablation) | U-Net · VOC 2012 | mIoU: 26.02 -> 30.42 | Improved |
| add CuMPerLay vectorization instead of SP PersLay (ablation) | TopoSwin-MP · ISIC | AUC: 95.60 -> 96.50 | Improved |
| add CuMPerLay vectorization instead of SP PersLay (ablation) | TopoSwin-SP · Glaucoma | AUC: 92.77 -> 92.90 | Improved |
Closest papers by meaning
- Conformable Convolution for Topologically Constrained Learning of Complex Anatomical Structures (AAAI 2026 · similarity 0.85)
- PLLay: Efficient Topological Layer based on Persistent Landscapes (NeurIPS 2020 · similarity 0.85)
- Towards Scalable Topological Regularizers (ICLR 2025 · similarity 0.85)
- Compositional PAC-Bayes: Generalization of GNNs with persistence and beyond (NeurIPS 2024 · similarity 0.83)
- Copresheaf Topological Neural Networks: A Generalized Deep Learning Framework (NeurIPS 2025 · similarity 0.83)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.82)
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.
