Data-centric AI building blocks for computer vision applications


Keywords
computer, vision, data, annotation, visualization, deep, learning, machine, computer-vision, data-annotation, data-visualization, deep-learning, machine-learning, python
License
CECILL-C
Install
pip install pixano==0.6.0b2

Documentation

Pixano

Data-centric AI building blocks for computer vision applications

Under active development, subject to API change

GitHub version PyPI version Tests Coverage Documentation Python version License


Pixano is an open-source tool by CEA List for exploring and annotating your dataset using AI features:

  • Fast dataset navigation using the the modern storage format Lance
  • Multi-view datasets support for images, and soon for 3D point clouds and videos
  • Import and export support for dataset formats like COCO
  • Semantic search using models like CLIP
  • Smart segmentation using models like SAM

Getting started

Installing Pixano

As Pixano requires specific versions for its dependencies, we recommend creating a new Python virtual environment to install it.

For example, with conda:

conda create -n pixano_env python=3.10
conda activate pixano_env

Then, you can install the Pixano package inside that environment with pip:

pip install pixano

Using your datasets

Please refer to our Jupyter notebooks for importing and exporting your datasets.

Using the Pixano app

Please refer to this link for using the Pixano app.

Contributing

Please refer to the CONTRIBUTING.md for information on running Pixano locally and guidelines on how to publish your contributions.

License

Pixano is licensed under the CeCILL-C license.