Research map / Probabilistic and scientific modeling / Variational and generative sampling
Bayesian Unsupervised Disentanglement of Anatomy and Geometry for Deep Groupwise Image Registration
Xinzhe Luo, Xin Wang, Linda G. Shapiro, Chun Yuan, Jianfeng Feng, Xiahai Zhuang
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 |
|---|---|---|---|
| train with complete groups instead of partial pairs (conv registration) (ablation) | Ours-PN · MS-CMRSeg | DSC: .803 -> .842 | Improved |
| train with complete groups instead of partial pairs (Demons registration) (ablation) | Ours-PD · MS-CMRSeg | DSC: .799 -> .836 | Improved |
| train with complete groups instead of partial pairs (ablation) | Ours-PN · BraTS-2021 | gWI: 0.608 -> 0.538 | Improved |
| train with complete groups instead of partial pairs (Demons) (ablation) | Ours-PD · BraTS-2021 | gWI: 0.720 -> 0.663 | Improved |
| use partial learning with group size 2 (ablation) | Ours-PN · MS-CMRSeg | DSC: .803 vs .799 | Mixed |
| replace convolutional registration with Demons registration (ablation) | Ours-CN · MS-CMRSeg | DSC: .842 -> .836 | No clear effect |
| replace convolutional registration with Demons registration (ablation) | Ours-CN · MS-CMRSeg | parameters (M, test): 5.06 -> 1.91 | Improved |
| replace convolutional registration with Demons registration (ablation) | Ours-CN · OASIS | negative Jacobian determinant voxels (%): .1746 -> .0066 | Improved |
| replace convolutional registration with Demons registration (ablation) | Ours-CN · BraTS-2021 | gWI: 0.538 -> 0.663 | Hurt |
| use proposed Bayesian disentangled model instead of best learning baseline | CTE-Att · MS-CMRSeg | DSC: .820 -> .842 | Improved |
| use proposed model instead of best learning baseline | APE-Att · Learn2Reg Abdominal MR-CT | DSC: .687 -> .803 | Improved |
| use proposed model instead of best baseline | APE · BraTS-2021 | DSC: .707 -> .709 | No clear effect |
Closest papers by meaning
- A Variational Bayesian Method for Similarity Learning in Non-Rigid Image Registration (CVPR 2022 · similarity 0.88)
- $\mathcal {X}$-Metric: An N-Dimensional Information-Theoretic Framework for Groupwise Registration and Deep Combined Computing (TPAMI 2022 · similarity 0.88)
- Uncertainty Estimation for Pretrained Medical Image Registration Models via Transformation Equivariance (NeurIPS 2026 · similarity 0.87)
- CARL: A Framework for Equivariant Image Registration (CVPR 2025 · similarity 0.86)
- Modality-Agnostic Structural Image Representation Learning for Deformable Multi-Modality Medical Image Registration (CVPR 2024 · similarity 0.85)
- Hierarchical Uncertainty Estimation for Learning-based Registration in Neuroimaging (ICLR 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.
