Research map / Detection and segmentation / CNN features
PMRNet: Physics-informed Multi-scale Refinement Network for Medical Image Segmentation
Boce Kang
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 Lightweight Memory Bank (ablation) | PMRNet · PH2 | HD95: 15.61 -> 13.99 | Improved |
| add Pseudo-Global Receptive Field (ablation) | PMRNet · PH2 | IoU: 90.43 -> 91.02 | Improved |
| add Boundary-aware Fusion decoder (ablation) | PMRNet · PH2 | IoU: 90.24 -> 91.02 | Improved |
| add all three modules LMB, PGRF, BF (ablation) | PMRNet · PH2 | HD95: 16.98 -> 13.99 | Improved |
| use PGRF instead of large-kernel conv (ablation) | PMRNet · Clinic | IoU: 86.14 -> 87.25 | Improved |
| use PGRF instead of dilated conv (ablation) | PMRNet · Clinic | IoU: 85.82 -> 87.25 | Improved |
| use PGRF instead of global pooling (ablation) | PMRNet · Clinic | IoU: 85.57 -> 87.25 | Improved |
| use ESConv instead of edge attention (ablation) | PMRNet · ISIC2017 | IoU: 82.21 -> 83.14 | Improved |
| use ESConv instead of Sobel filter + Conv (ablation) | PMRNet · ISIC2017 | IoU: 81.87 -> 83.14 | Improved |
| use ESConv instead of Canny edge + Conv (ablation) | PMRNet · ISIC2017 | IoU: 81.63 -> 83.14 | Improved |
| increase progressive refinement iterations beyond 3 (ablation) | PMRNet · PH2 | IoU | No clear effect |
| PMRNet versus larger models on DDTI | PMRNet · DDTI | IoU: 82.52 vs Deformable-LKA 82.81 | Mixed |
Closest papers by meaning
- Fractal-G: Topology-Aware Heterogeneous Graphs for Medical Image Segmentation (NeurIPS 2026 · similarity 0.85)
- Modality-agnostic Domain Generalizable Medical Image Segmentation by Multi-Frequency in Multi-Scale Attention (CVPR 2024 · similarity 0.85)
- Scale-Space Hypernetworks for Efficient Biomedical Image Analysis (NeurIPS 2023 · similarity 0.84)
- DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.84)
- Knowledge Transfer Scaling Laws for 3D Medical Imaging (NeurIPS 2026 · similarity 0.84)
- GapMatch: Bridging Instance and Model Perturbations for Enhanced Semi-Supervised Medical Image Segmentation (AAAI 2025 · similarity 0.83)
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.
