Research map / Detection and segmentation / Open-set and few-shot detection
UKBOB: One Billion MRI Labeled Masks for Generalizable 3D Medical Image Segmentation
Emmanuelle Bourigault, Amir Jamaludin, Abdullah Hamdi
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 SOLF label filtering to UKBOB training labels (ablation) | Swin-BOB · AMOS zero-shot | mean Dice: 0.818 -> 0.840 | Improved |
| add SOLF label filtering to UKBOB training labels (ablation) | Swin-BOB · BTCV zero-shot | mean Dice: 0.856 -> 0.882 | Improved |
| use full SOLF (volume, eccentricity, sphericity) instead of volume-only filter (ablation) | Swin-BOB · BTCV zero-shot | mean Dice: 0.875 -> 0.882 | Improved |
| use full SOLF instead of volume-only filter (ablation) | Swin-BOB · AMOS zero-shot | mean Dice: 0.832 -> 0.840 | Improved |
| apply SOLF filtering (epsilon=2) to spine labels versus no filtering (ablation) | UKBOB labels · Spine manual labels | Dice: 0.811 -> 0.867 | Improved |
| apply SOLF filtering (epsilon=2) to abdomen labels versus no filtering (ablation) | UKBOB labels · UKBOB-manual | Dice: 0.873 -> 0.891 | Improved |
| use SOLF instead of IQR filtering (ablation) | UKBOB labels · Spine manual labels | Dice: 0.849 -> 0.867 | Improved |
| vary SOLF threshold epsilon from 3 to 5 (ablation) | Swin-UNetr · UKBOB-manual abdomen | mean Dice: 0.891 -> 0.745 | Hurt |
| add ETTA to nn-UNet | nn-UNet · BTCV | mean Dice: 0.804 -> 0.831 | Improved |
| add ETTA to Swin-UNetr | Swin-UNetr · AMOS | mean Dice: 0.822 -> 0.858 | Improved |
| add ETTA to Swin-BOB | Swin-BOB · BRATS | mean Dice: 0.882 -> 0.894 | Improved |
| replace standard TTA with ETTA | Swin-BOB · AMOS | mean Dice: 0.857 -> 0.864 | Improved |
Closest papers by meaning
- AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation (NeurIPS 2022 · similarity 0.87)
- Mine yOur owN Anatomy: Revisiting Medical Image Segmentation With Extremely Limited Labels (TPAMI 2024 · similarity 0.87)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (CVPR 2024 · similarity 0.86)
- VISTA3D: A Unified Segmentation Foundation Model For 3D Medical Imaging (CVPR 2025 · similarity 0.85)
- Better Tokens for Better 3D: Advancing Vision-Language Modeling in 3D Medical Imaging (NeurIPS 2025 · similarity 0.85)
- Unleashing the Potential of SAM for Medical Adaptation via Hierarchical Decoding (CVPR 2024 · 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.
