Research map / Label-efficient and robust learning / Semi-supervised learning
One-Shot Medical Landmark Localization by Edge-Guided Transform and Noisy Landmark Refinement
Zihao Yin, Ping Gong, Chunyu Wang, Yizhou Yu, Yizhou Wang
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 learned global affine alignment (ablation) | registration stage I · Head | MRE (mm): 3.42 -> 2.70 | Improved |
| learned affine instead of SIFT alignment (ablation) | registration stage I · Head | MRE (mm): 3.10 -> 2.70 | Improved |
| add a second local deformation step (ablation) | registration stage I · Head | MRE (mm): 2.73 -> 2.70 | Improved |
| use perspective instead of affine transform (ablation) | registration stage I · Head | MRE (mm): 2.70 -> 2.80 | Hurt |
| add attention blocks and multi-head attention (ablation) | registration stage I · Head | MRE (mm): 2.93 -> 2.70 | Improved |
| HRNet instead of UNet backbone (ablation) | registration stage I · Head | MRE (mm): 2.95 -> 2.70 | Improved |
| add invertibility loss L_inv (ablation) | registration stage I · Head | MRE (mm): 3.24 -> 2.70 | Improved |
| add smoothness loss L_smooth (ablation) | registration stage I · Head | MRE (mm): 2.97 -> 2.70 | Improved |
| add synthetic-pair loss L_syn (ablation) | registration stage I · Head | MRE (mm): 2.86 -> 2.70 | Improved |
| edge-masked similarity L_esim instead of L_sim (ablation) | registration stage I · Head | MRE (mm): 3.17 -> 2.70 | Improved |
| edge-aware smoothness L_esmooth instead of L_smooth (ablation) | registration stage I · Head | MRE (mm): 2.75 -> 2.70 | Improved |
| add self-consistency filtering L_filter (ablation) | C2T stage II · Head | MRE (mm): 2.53 -> 2.29 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- 3FabRec — 3FabRec: Fast Few-Shot Face Alignment by Reconstruction (CVPR 2020) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Improving Landmark Localization With Semi-Supervised Learning (CVPR 2018 · similarity 0.88)
- Domain Generalized Medical Landmark Detection via Robust Boundary-Aware Pre-Training (AAAI 2025 · similarity 0.88)
- MCPNet:Masked Coordinate Pooling-based Attention Network for Medical Landmark Detection (ECCV 2026 · similarity 0.85)
- Multi-Domain Multi-Definition Landmark Localization for Small Datasets (ECCV 2022 · similarity 0.85)
- FaceLift: Semi-supervised 3D Facial Landmark Localization (CVPR 2024 · similarity 0.84)
- POPoS: Improving Efficient and Robust Facial Landmark Detection with Parallel Optimal Position Search (AAAI 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.
