[ANN] PointCloudRegistration.jl

Package Announcements
1

Hi everyone!
I’d like to announce a new package I developed, called PointCloudRegistration.jl.
It’s purpose is to rigidly (i.e. rotate, translate, and reflect) and non-rigidly (i.e. deform) register pairs of point clouds. I mainly use it for applications in structural biology but it is intended to be fairly domain agnostic!

It’s currently in the process of being added to the General Package Registry and I will present it at JuliaCon next Wednesday :slight_smile: So see you there if you are interested in details!

I cannot post links here, so you can check out the package on Codeberg, user a5s, repo PointCloudRegistration.jl.

2

here is the link (or i hope its the link)

3

Hi, I have an interest in unsupervised medical image registration (ie KeyMorph [2304.09941] A Robust and Interpretable Deep Learning Framework for Multi-modal Registration via Keypoints), and the idea of registering point-clouds made me think of it again, especially when it comes to doing it in a differentiable way without standard SVD. I’m already going through the different algorithms looking for new ways to frame my problem and am looking forward to your talk!