Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Modeling the Probabilistic Distribution of Unlabeled Data for One-shot Medical Image Segmentation
Read the paperCheck 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 probabilistic shape and intensity VAEs to generate deformations | One-shot brain MRI segmentation · CANDI test | Mean Dice (%): 82.3 -> 85.1 | Helped |
| Use probabilistic shape and intensity VAEs to generate deformations | One-shot brain MRI segmentation · ABIDE seen test | Mean Dice (%): 76.7 vs 70.3; 76.7 vs 69.6 | Helped |
| Use probabilistic shape and intensity VAEs to generate deformations | One-shot brain MRI segmentation · ABIDE unseen test | Mean Dice (%): 74.8 vs 64.3; 74.8 vs 62.9 | Helped |
| Add VAE-generated shape deformations (ablation) | Shape-augmented segmentation · CANDI test | Mean Dice (%): 81.7 -> 83.5 | Helped |
| Add registered intensity deformations (ablation) | Shape-augmented segmentation · CANDI test | Mean Dice (%): 81.7 -> 84.2 | Helped |
| Add VAE-generated intensity deformations alongside shape augmentation (ablation) | Shape-augmented segmentation · CANDI test | Mean Dice (%): 83.5 -> 85.1 | Helped |
| Add the warped-image intensity reconstruction loss to the shape VAE (ablation) | Shape VAE using deformation reconstruction loss · CANDI test | Mean Dice (%): 81.3 -> 83.5 | Helped |
| Add the deformation reconstruction loss to the shape VAE (ablation) | Shape VAE using image-intensity reconstruction loss · CANDI test | Mean Dice (%): 82.3 -> 83.5 | Helped |
| Use smaller beta in the VAE KL-divergence objective (ablation) | Shape and intensity VAEs · CANDI test | Segmentation accuracy | Helped |
| Sample latent codes with standard deviation sigma 1 rather than a larger sigma (ablation) | Shape and intensity VAEs · CANDI test | Segmentation performance | Hurt |
| Increase latent sampling standard deviation from 10 to 100 (ablation) | Shape and intensity VAEs · CANDI test | Segmentation accuracy | No clear effect |
Closest papers by meaning
- Data Augmentation Using Learned Transformations for One-Shot Medical Image Segmentation (CVPR 2019 · similarity 0.91)
- LT-Net: Label Transfer by Learning Reversible Voxel-Wise Correspondence for One-Shot Medical Image Segmentation (CVPR 2020 · similarity 0.88)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (CVPR 2024 · similarity 0.86)
- One-Prompt to Segment All Medical Images (CVPR 2024 · similarity 0.85)
- Revisiting MAE Pre-training for 3D Medical Image Segmentation (CVPR 2025 · similarity 0.85)
- Combinatorial CNN-Transformer Learning with Manifold Constraints for Semi-supervised Medical Image Segmentation (AAAI 2024 · 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.