Research map / Detection and segmentation / CNN features
TopoAgent: An Agentic Framework for Automated Topology Learning in Medical Imaging
Guangyu Meng, Xiaowei Xu, Xueyang Li, Yiyu Shi, Erin Chambers, Danny 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 TopoAgent's full pipeline versus GPT-4o with the same tools in a ReAct loop | GPT-4o · TopoBenchmark avg | balanced accuracy (%): 46.25 -> 68.21 | Improved |
| add per-image adaptive descriptor determination over a fixed per-object-type oracle | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 58.89 -> 68.21 | Improved |
| add the skill set S (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 59.56 -> 68.21 | Improved |
| add Perception phase (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 61.22 -> 68.21 | Improved |
| add Reflection phase (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 61.68 -> 68.21 | Improved |
| add dual memory (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 62.90 -> 68.21 | Improved |
| add independent proposal step before seeing rankings (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 59.76 -> 68.21 | Improved |
| add S_rank (empirical rankings) (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 60.93 -> 68.21 | Improved |
| add S_param and S_prop (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 62.63 -> 68.21 (S_param); 64.11 -> 68.21 (S_prop) | Improved |
| add PH profile and visual statistics to perception (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 64.12 -> 68.21 (PH profile); 64.95 -> 68.21 (visual stats) | Improved |
| add short-term memory (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 66.73 -> 68.21 | Improved |
| replace GPT-4o backbone with Claude 4.6 or Gemini 2.5 Pro (ablation) | TopoAgent · TopoBenchmark avg | balanced accuracy (%): 68.21 -> 68.37 (Claude 4.6); 67.98 (Gemini 2.5 Pro) | No clear effect |
Closest papers by meaning
- TopoCL: Topological Contrastive Learning for Medical Imaging (CVPR 2026 · similarity 0.92)
- Topology-Reinforced Swin Transformer for Medical Image Analysis (NeurIPS 2026 · similarity 0.90)
- TopoSlide: Topologically-Informed Histopathology Whole Slide Image Representation Learning (CVPR 2026 · similarity 0.89)
- TopoDiffusionNet: A Topology-aware Diffusion Model (ICLR 2025 · similarity 0.88)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.86)
- TopoCellGen: Generating Histopathology Cell Topology with a Diffusion Model (CVPR 2025 · 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.
