In this talk, we will introduce Kornia-rs, an Open Source Computer Vision and 3D Library developed natively in Rust. Kornia-rs facilitates low-level image and 3D data operations essential for machine learning and data science pipelines. The library leverages Rust's thread safety and efficiency, making it highly suitable for robotics and production environments.

Rust is here to stay and is rapidly growing in interest among the Robotics and Machine Learning communities, becoming one of the first alternative programming language choices. Much of the existing tooling is being rewritten in Rust, and Kornia-rs is no exception. It is a fresh port of the popular Kornia library in Pytorch, initially created as an alternative to OpenCV for performing image geometric transformations within the Pytorch community and now supported by other frameworks such as Numpy, TensorFlow, and Jax.

During the session, we'll explore the history and evolution of the library where we will cover some of the basics, and showcase how Kornia-rs can be seamlessly integrated with existing tools like Rerun or Zenoh to build efficient data pipeline applications for Robotics.

Link to the project: https://github.com/kornia/kornia-rs