[NeurIPS 2025] Deformable 2D/3D registration via polyrigid transforms
-
Updated
Oct 22, 2025 - Python
[NeurIPS 2025] Deformable 2D/3D registration via polyrigid transforms
Official code for GPOreg (MICCAI 2025, early accept) — Gaussian Primitive Optimized deformable retinal image registration.
[Medical Image Analysis] Residual Aligner-based Network (RAN): Motion-Aware Structure for Coarse-to-fine Discontinuous Deformable Registration
PyTorch implementation of the descriptor DEAL presented at NeurIPS 2021 "Extracting Deformation-Aware Local Features by Learning to Deform".
Light-weight Deformable Registration using Adversarial Learning with Distilling Knowledge (IEEE Transactions on Medical Imaging 2021))
PyTorch framework for biomechanically constrained deformable image registration with support for locally rigid structures, sliding interfaces, and learned tissue mechanics. Associated paper: https://arxiv.org/abs/2504.05444
Probabilistic deformable medical image registration capable of uncertainty quantification using a hierarchical variational autoencoder (VAE) architecture. Published at MICCAI 2024 conference.
Respiratory motion forecasting from thoraco-abdominal cine MRI and external-marker position time series using RNNs trained online and transformers
🧠 MATLAB toolbox for grayscale image registration using affine control points, optimization, and deformable methods.
MICCAI'2025 | VoxelOpt: training-free deformable abdominal CT registration with voxel-adaptive message passing. Codex and Claude Code friendly PyTorch repo.
Demo of deformable registration methods utilizing voxelmorph
Reproducible computational workflows for biomedical imaging, deformable registration, uncertainty quantification, and medical physics research.
Landmark-initialized diffeomorphic pipeline for CT–MRI liver registration (ablation planning) Research — documentation only, code confidential.
Python implementation of training a model for multimodal deformable image registration
Intensity-based deformable CT–CT liver registration pipeline (contrast-enhanced to non-contrast). Research — documentation only, code confidential
To associate your repository with the deformable-registration topic, visit your repo's landing page and select "manage topics."