Welcome to GeomFuM

GeomFuM is a Modular Python Package for Machine Learning with Functional Maps.

Functional Maps provide a powerful framework for representing and computing correspondences between geometric shapes. GeomFuM offers a comprehensive suite of tools for:

• Shape Analysis: Laplace-Beltrami eigenfunctions, heat kernel signatures, wave kernel signatures

• Correspondence Computation: Functional maps, pointwise correspondences, refinement algorithms

• Deep Learning Integration: Neural functional maps, feature learning for shape correspondence

• Geometric Processing: Mesh operations, hierarchical representations, sampling strategies

Installation

pip install geomfum@git+https://github.com/DiG-AIR/geomfum.git@main

Contents

• Installation
• Usage
• Concepts
• Tutorials
• Contributing
• API Reference

Indices and tables

• Index

• Module Index

• Search Page