Research map / Architectures and efficiency / Neural architecture search
C2FNAS: Coarse-to-Fine Neural Architecture Search for 3D Medical Image Segmentation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use a 5-fold ensemble of C2FNAS-Panc | C2FNAS-Panc · MSD (Avg Task) | DSC: 75.87 -> 76.97 | Helped |
| Use C2FNAS-Panc* instead of nnU-Net | MSD · MSD (Avg Task) | DSC: 76.39 -> 76.97 | Helped |
| Search cell operations after topology search rather than using only standard 3D operations (ablation) | C2FNAS · Pancreas (Avg, 5-fold validation) | DSC: 66.86 -> 68.37 | Helped |
| Search cell operations after topology search rather than using only standard 3D operations (ablation) | C2FNAS · Lung (5-fold validation) | DSC: 69.05 -> 69.77 | Helped |
| Search topology on Lung rather than Pancreas (ablation) | C2FNAS coarse stage · Lung (5-fold validation) | DSC: 69.05 -> 71.74 | Helped |
| Search topology on Pancreas rather than Lung (ablation) | C2FNAS coarse stage · Pancreas (Avg, 5-fold validation) | DSC: 66.39 -> 66.86 | Helped |
| Scale filter count across the tested range (ablation) | C2FNAS-Panc · Lung, Pancreas, and Hippocampus (single-fold validation) | DSC: 72.32; 73.89; 76.15; 74.26; 76.94; 75.37; 75.98; 77.75 | Mixed |
| Increase filter scaling from 1.00 to 2.00 (ablation) | C2FNAS-Panc · Pancreas (Avg, single-fold validation) | DSC: 65.34 -> 66.40 | Helped |
| Increase filter scaling from 1.00 to 2.00 (ablation) | C2FNAS-Panc · Hippocampus (Avg, single-fold validation) | DSC: 81.96 -> 82.46 | Helped |
Closest papers by meaning
- DiNTS: Differentiable Neural Network Topology Search for 3D Medical Image Segmentation (CVPR 2021 · similarity 0.91)
- HyperSegNAS: Bridging One-Shot Neural Architecture Search With 3D Medical Image Segmentation Using HyperNet (CVPR 2022 · similarity 0.91)
- Upping the Game: How 2D U-Net Skip Connections Flip 3D Segmentation (NeurIPS 2024 · similarity 0.87)
- E2ENet: Dynamic Sparse Feature Fusion for Accurate and Efficient 3D Medical Image Segmentation (NeurIPS 2024 · similarity 0.86)
- Revisiting MAE Pre-training for 3D Medical Image Segmentation (CVPR 2025 · similarity 0.86)
- EffiDec3D: An Optimized Decoder for High-Performance and Efficient 3D Medical Image Segmentation (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.