Research map / Vision-language and multimodal / Cross-modal learning
TopoCL: Topological Contrastive Learning for Medical Imaging
Guangyu Meng, Pengfei Gu, Peixian Liang, John P. Lalor, Erin Wolf Chambers, Danny Z. 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 |
|---|---|---|---|
| add pretraining of both encoders (ablation) | SimCLR+TopoCL · PathMNIST | accuracy: 68.98 -> 93.21 | Improved |
| add visual encoder pretraining (ablation) | SimCLR+TopoCL · PathMNIST | accuracy: 81.53 -> 93.21 | Improved |
| add topology encoder pretraining (ablation) | SimCLR+TopoCL · PathMNIST | accuracy: 87.64 -> 93.21 | Improved |
| add hierarchical attention in topology encoder (ablation) | SimCLR+TopoCL · PathMNIST | accuracy: 86.51 -> 93.21 | Improved |
| add cross-attention between H0 and H1 (ablation) | SimCLR+TopoCL · PathMNIST | accuracy: 87.59 -> 93.21 | Improved |
| add self-attention within each homology dimension (ablation) | SimCLR+TopoCL · PathMNIST | accuracy: 91.24 -> 93.21 | Improved |
| add cross-attention fusion expert (ablation) | SimCLR+TopoCL · PathMNIST/OrganS/OCT | accuracy: 1.73% average drop when removed | Improved |
| add gated blending expert (ablation) | SimCLR+TopoCL · PathMNIST/OrganS/OCT | accuracy: 1.45% average drop when removed | Improved |
| add visual-only expert (ablation) | SimCLR+TopoCL · PathMNIST/OrganS/OCT | accuracy: 1.01% average drop when removed | Improved |
| add topology-only expert (ablation) | SimCLR+TopoCL · PathMNIST/OrganS/OCT | accuracy: 0.41% average drop when removed | Improved |
| use topology-aware augmentations instead of standard visual ones (ablation) | MoCo-v3+TopoCL · OrganSMNIST | accuracy: 75.51 -> 80.58 | Improved |
| compute PDs on SAM-extracted ROIs instead of full image (ablation) | MoCo-v3+TopoCL · OrganSMNIST | accuracy: 76.34 -> 80.58 | Improved |
Closest papers by meaning
- TopoAgent: An Agentic Framework for Automated Topology Learning in Medical Imaging (ECCV 2026 · similarity 0.92)
- Topology-Reinforced Swin Transformer for Medical Image Analysis (NeurIPS 2026 · similarity 0.89)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.85)
- TopoGCL: Topological Graph Contrastive Learning (AAAI 2024 · similarity 0.85)
- TopoDiffusionNet: A Topology-aware Diffusion Model (ICLR 2025 · similarity 0.85)
- TopoCellGen: Generating Histopathology Cell Topology with a Diffusion Model (CVPR 2025 · similarity 0.84)
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.
