amosgropp/IGR
A deep learning method that learns implicit signed distance representations from raw point clouds for 3D surface reconstruction.
Not currently ranked — collecting fresh signals.
star history
This repository implements a deep learning approach for reconstructing 3D surfaces from point cloud data. The method optimizes a neural network to solve the eikonal equation, using the input point cloud as a boundary condition to find a signed distance function (SDF). The implicit geometric regularization guides the optimization toward natural solutions for shape representation. The implementation uses PyTorch and supports various point cloud formats including xyz, npy, npz, and ply files.
Frequently asked
- What is amosgropp/IGR?
- A deep learning method that learns implicit signed distance representations from raw point clouds for 3D surface reconstruction.
- Is IGR open source?
- Yes — amosgropp/IGR is an open-source project tracked on heatdrop.
- What language is IGR written in?
- amosgropp/IGR is primarily written in Python.
- How popular is IGR?
- amosgropp/IGR has 447 stars on GitHub.
- Where can I find IGR?
- amosgropp/IGR is on GitHub at https://github.com/amosgropp/IGR.