README.md

Event Vision Library

Installation

You can install Event Vision Library via pip from PyPI:

$ pip install event-vision-library

# you can now `import evlib`

Usage

Please see our examples and documentation.

Features

• Python 3.7, 3.8, 3.9, 3.10
• Pure-python library
• Numpy and Torch compatibility.
• 🚧 This library is under construction and currently alpha version. The APIs may change significantly. Contributions and discussions are welcomed! 🚧

Data

• Support different data types (.text, .raw, .hdf5, .npy, .aedat) for various file encoding of event data
• ROS bag files (optional, based on ROS installation)
• Support multiple existing dataset (e.g., ECD, MVSEC, DSEC, etc.)
• Support iterator-based loading and also block-based (random access) loading.

Algorithms

• Have different off-the-shelf methods, ready to use:
  • Optical Flow estimation
  • Image reconstruction
  • Ego-motion estimation
  • more to come.
• C++ implementation and extension for faster execution (TODO)

Log and Vsualization

• Various visualization for 2D/3D representation of events
• Useful logging

Contributing

Contributions are very welcome. To learn more, see the Contributor Guide.

License

Distributed under the terms of the MIT license, Event Vision Library is free and open source software.

Issues

If you encounter any problems, please file an issue along with a detailed description.

Acknowledgement

This project was generated from Hypermodern Python Cookiecutter template.