https://github.com/brentyi/jaxlie

Rigid transforms + Lie groups in JAX
https://github.com/brentyi/jaxlie

computer-vision geometry jax lie-groups robotics

Last synced: 4 days ago
JSON representation

Rigid transforms + Lie groups in JAX

• Host: GitHub
• URL: https://github.com/brentyi/jaxlie
• Owner: brentyi
• License: mit
• Created: 2020-11-28T06:18:13.000Z (about 4 years ago)
• Default Branch: master
• Last Pushed: 2024-11-30T05:56:23.000Z (about 1 month ago)
• Last Synced: 2024-12-28T07:04:22.825Z (11 days ago)
• Topics: computer-vision, geometry, jax, lie-groups, robotics
• Language: Python
• Homepage:
• Size: 13.3 MB
• Stars: 237
• Watchers: 7
• Forks: 15
• Open Issues: 6
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-jax - jaxlie - Lie theory library for rigid body transformations and optimization. <img src="https://img.shields.io/github/stars/brentyi/jaxlie?style=social" align="center"> (Libraries / New Libraries)

README

        
# jaxlie

![build](https://github.com/brentyi/jaxlie/workflows/build/badge.svg)

![mypy](https://github.com/brentyi/jaxlie/workflows/mypy/badge.svg)

![lint](https://github.com/brentyi/jaxlie/workflows/lint/badge.svg)

[![codecov](https://codecov.io/gh/brentyi/jaxlie/branch/master/graph/badge.svg)](https://codecov.io/gh/brentyi/jaxlie)

[![pypi_dowlnoads](https://pepy.tech/badge/jaxlie)](https://pypi.org/project/jaxlie)

**[ [API reference](https://brentyi.github.io/jaxlie) ]** **[

[PyPI](https://pypi.org/project/jaxlie/) ]**

`jaxlie` is a library containing implementations of Lie groups commonly used for

rigid body transformations, targeted at computer vision & robotics

applications written in JAX. Heavily inspired by the C++ library

[Sophus](https://github.com/strasdat/Sophus).

We implement Lie groups as high-level (data)classes:

  

    

      Group

      Description

      Parameterization

    

  

  

    

      jaxlie.SO2

      Rotations in 2D.

      (real, imaginary): unit complex (∈ S1)

    

    

      jaxlie.SE2

      Proper rigid transforms in 2D.

      (real, imaginary, x, y): unit complex & translation

    

    

      jaxlie.SO3

      Rotations in 3D.

      (qw, qx, qy, qz): wxyz quaternion (∈ S3)

    

    

      jaxlie.SE3

      Proper rigid transforms in 3D.

      (qw, qx, qy, qz, x, y, z): wxyz quaternion & translation

    

  

Where each group supports:

- Forward- and reverse-mode AD-friendly **`exp()`**, **`log()`**,

  **`adjoint()`**, **`apply()`**, **`multiply()`**, **`inverse()`**,

  **`identity()`**, **`from_matrix()`**, and **`as_matrix()`** operations. (see

  [./examples/se3_example.py](./examples/se3_basics.py))

- Taylor approximations near singularities.

- Helpers for optimization on manifolds (see

  [./examples/se3_optimization.py](./examples/se3_optimization.py),

  jaxlie.manifold.\*).

- Compatibility with standard JAX function transformations. (see

  [./examples/vmap_example.py](./examples/vmap_example.py))

- Broadcasting for leading axes.

- (Un)flattening as pytree nodes.

- Serialization using [flax](https://github.com/google/flax).

We also implement various common utilities for things like uniform random

sampling (**`sample_uniform()`**) and converting from/to Euler angles (in the

`SO3` class).

---

### Install (Python >=3.7)

```bash

# Python 3.6 releases also exist, but are no longer being updated.

pip install jaxlie

```

---

### Misc

`jaxlie` was originally written when I was learning about Lie groups for our IROS 2021 paper

([link](https://github.com/brentyi/dfgo)):

```

@inproceedings{yi2021iros,

    author={Brent Yi and Michelle Lee and Alina Kloss and Roberto Mart\'in-Mart\'in and Jeannette Bohg},

    title = {Differentiable Factor Graph Optimization for Learning Smoothers},

    year = 2021,

    BOOKTITLE = {2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)}

}

```