ranahanocka/point2mesh
A deep learning technique that reconstructs watertight 3D meshes from point cloud data by optimizing CNN weights to deform an initial mesh.
Not currently ranked — collecting fresh signals.
star history
Point2Mesh is a SIGGRAPH 2020 technique for surface mesh reconstruction from point clouds. It uses a CNN trained on a single object to learn local-scale geometric self-similarity, optimizing convolutional kernel weights to shrink-wrap an initial mesh onto the input point cloud. The implementation uses PyTorch and PyTorch3D for deep learning operations.
Frequently asked
- What is ranahanocka/point2mesh?
- A deep learning technique that reconstructs watertight 3D meshes from point cloud data by optimizing CNN weights to deform an initial mesh.
- Is point2mesh open source?
- Yes — ranahanocka/point2mesh is open source, released under the MIT license.
- What language is point2mesh written in?
- ranahanocka/point2mesh is primarily written in Python.
- How popular is point2mesh?
- ranahanocka/point2mesh has 1.2k stars on GitHub.
- Where can I find point2mesh?
- ranahanocka/point2mesh is on GitHub at https://github.com/ranahanocka/point2mesh.