Research map / Detection and segmentation / Open-set and few-shot detection
Multi-Atlas Segmentation Using Partially Annotated Data: Methods and Annotation Strategies
Lisa Margret Koch, Martin Rajchl, Wenjia Bai, Christian Frederik Baumgartner, Tong Tong, Jonathan Passerat‐Palmbach, Paul Aljabar, Daniel Rueckert
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 spatial regularisation to locally weighted label fusion (MASr-LW) (ablation) | MAS-LW · ADNI hippocampus, R=5,10,15,20 atlases | mean Dice | Improved |
| use locally weighted fusion instead of majority vote | multi-atlas segmentation · ADNI hippocampus, 5 or 10 atlases | mean Dice | Hurt |
| use locally weighted fusion instead of majority vote | multi-atlas segmentation · ADNI hippocampus, larger atlas sets | mean Dice | Improved |
| increase number of atlases | multi-atlas segmentation · ADNI hippocampus | mean Dice | Improved |
| reduce labelled atlas slices to q=0.4 (ablation) | PA-SW-CONF1 and CONF2 · ADNI hippocampus | mean Dice | No clear effect |
| reduce labelled atlas slices below q=0.4 (ablation) | PA-SW-CONF1 · ADNI hippocampus | mean Dice | Hurt |
| add inter-atlas label propagation (CONF2 instead of CONF1) (ablation) | PA-SW · ADNI hippocampus, q small | mean Dice: 0.83 ± 0.08 at q=0.1 for CONF2 | Improved |
| use scribbled atlases instead of fully annotated atlases | MASr-LW · cardiac MR, right and left ventricle | mean Dice | No clear effect |
| use scribbled atlases instead of fully annotated atlases | MASr-LW · cardiac MR, myocardium | mean Dice | Hurt |
| add scribbled atlases to scribbled target (ablation) | PA-SC-T · cardiac MR | mean Dice | No clear effect |
| add target scribbles to fully annotated atlases | MASr-LW · cardiac MR | mean Dice | Improved |
| scribble annotation instead of full annotation | annotation effort · cardiac MR | annotation time: 30 min -> 3.9 ± 0.6 min | Improved |
Closest papers by meaning
- Alternative Baselines for Low-Shot 3D Medical Image Segmentation—An Atlas Perspective (AAAI 2021 · similarity 0.89)
- Disentangling Human Error from Ground Truth in Segmentation of Medical Images (NeurIPS 2020 · similarity 0.87)
- Learning to Segment from Noisy Annotations: A Spatial Correction Approach (ICLR 2023 · similarity 0.87)
- Annotation Ambiguity Aware Semi-Supervised Medical Image Segmentation (CVPR 2025 · similarity 0.86)
- SGTC: Semantic-Guided Triplet Co-training for Sparsely Annotated Semi-Supervised Medical Image Segmentation (AAAI 2025 · similarity 0.85)
- Pancakes: Consistent Multi-Protocol Image Segmentation Across Biomedical Domains (NeurIPS 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.
