Reproduce “BINDER: A Latent Variable Model for Probabilistic Medical Image Registration” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Stefano Cerri, Amirhossein Hassankhani, Yaël Balbastre, Koen Van Leemput
BINDER models voxel correspondence in multimodal images as a latent variable, weighting nearby candidates by B-spline proximity and intensity compatibility. The model supports an EM registration optimizer and a Gibbs sampler for alternative deformations. Across nine tasks, it reached Dice 0.739 versus 0.336 for Elastix in abdominal MR-to-CT registration with transferred configurations. Regularization remains user-set; sampling slow and assessment limited.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.