Skip to main content

Getting Started With ADT

Overview​

This section provides a step-by-step guide to run the Aria Digital Twin (ADT) quickstart tutorial in a Jupyter notebook.

This notebook provides a walkthrough of:

  • Loading an ADT sequence
  • Accessing and visualizing all ADT ground-truth data:
    • 6DoF object poses
    • 2D object bounding box
    • Segmentation images
    • Depth images
    • Skeleton and synthetic rendering
    • Eye gaze
  • An example of undistorting ADT ground-truth data

We also have an two Advanced Tutorials that will walk you through getting synchronized ground truth data in a multi-person sequence, and using depth maps + RGB images to generated 3D colored pointclouds.

Run Jupyter Notebook on Google Colab​

Use the following link to run the Python notebook in an installation free playground:

Running the Jupyter Notebook locally​

Step 0 : Check system requirements and download codebase​

Ensure your system is supported and then download projectaria_tools codebase from the github

Step 1 : Install Python​

If you have already installed projectaria-tools using Python Package Installation, you can skip to Step 4. The ADT Python code is part of the main projectaria-tools package.

Jupyter notebook error

If you have problems using Jupyter examples, please upgrade python3 to the latest version.

Step 2 : Create a virtual environment​

rm -rf $HOME/projectaria_tools_python_env

python3 -m venv $HOME/projectaria_tools_python_env

source $HOME/projectaria_tools_python_env/bin/activate

Step 3 : Install projectaria_tools from pypi​

pip3 install --upgrade pip

pip3 install projectaria-tools'[all]'

Step 4 : Download Sample Sequence:​

Download the sample ADT sequence by following steps 0 to 4 in How to Download the ADT Dataset.

Step 5 : Run Tutorial​

From your projectaria_tools python virtual environment, run:

cd $HOME/Documents/projectaria_sandbox

jupyter notebook projectaria_tools/projects/AriaDigitalTwinDatasetTools/examples/adt_quickstart_tutorial.ipynb

Troubleshooting​

Go to Data Utilities Troubleshooting if you have issues implementing this guide.